aboutsummaryrefslogtreecommitdiff
path: root/tools/power/x86/turbostat/turbostat.c
diff options
context:
space:
mode:
Diffstat (limited to 'tools/power/x86/turbostat/turbostat.c')
-rw-r--r--tools/power/x86/turbostat/turbostat.c1380
1 files changed, 1094 insertions, 286 deletions
diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c
index 2655ae9a3ad..d0396af99fa 100644
--- a/tools/power/x86/turbostat/turbostat.c
+++ b/tools/power/x86/turbostat/turbostat.c
@@ -2,7 +2,7 @@
* turbostat -- show CPU frequency and C-state residency
* on modern Intel turbo-capable processors.
*
- * Copyright (c) 2012 Intel Corporation.
+ * Copyright (c) 2013 Intel Corporation.
* Len Brown <len.brown@intel.com>
*
* This program is free software; you can redistribute it and/or modify it
@@ -20,7 +20,10 @@
*/
#define _GNU_SOURCE
+#include MSRHEADER
+#include <stdarg.h>
#include <stdio.h>
+#include <err.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/wait.h>
@@ -34,30 +37,26 @@
#include <string.h>
#include <ctype.h>
#include <sched.h>
-
-#define MSR_NEHALEM_PLATFORM_INFO 0xCE
-#define MSR_NEHALEM_TURBO_RATIO_LIMIT 0x1AD
-#define MSR_IVT_TURBO_RATIO_LIMIT 0x1AE
-#define MSR_APERF 0xE8
-#define MSR_MPERF 0xE7
-#define MSR_PKG_C2_RESIDENCY 0x60D /* SNB only */
-#define MSR_PKG_C3_RESIDENCY 0x3F8
-#define MSR_PKG_C6_RESIDENCY 0x3F9
-#define MSR_PKG_C7_RESIDENCY 0x3FA /* SNB only */
-#define MSR_CORE_C3_RESIDENCY 0x3FC
-#define MSR_CORE_C6_RESIDENCY 0x3FD
-#define MSR_CORE_C7_RESIDENCY 0x3FE /* SNB only */
+#include <cpuid.h>
char *proc_stat = "/proc/stat";
unsigned int interval_sec = 5; /* set with -i interval_sec */
unsigned int verbose; /* set with -v */
-unsigned int summary_only; /* set with -s */
+unsigned int rapl_verbose; /* set with -R */
+unsigned int rapl_joules; /* set with -J */
+unsigned int thermal_verbose; /* set with -T */
+unsigned int summary_only; /* set with -S */
+unsigned int dump_only; /* set with -s */
unsigned int skip_c0;
unsigned int skip_c1;
unsigned int do_nhm_cstates;
unsigned int do_snb_cstates;
+unsigned int do_c8_c9_c10;
+unsigned int do_slm_cstates;
+unsigned int use_c1_residency_msr;
unsigned int has_aperf;
-unsigned int units = 1000000000; /* Ghz etc */
+unsigned int has_epb;
+unsigned int units = 1000000; /* MHz etc */
unsigned int genuine_intel;
unsigned int has_invariant_tsc;
unsigned int do_nehalem_platform_info;
@@ -67,6 +66,7 @@ unsigned int extra_msr_offset32;
unsigned int extra_msr_offset64;
unsigned int extra_delta_offset32;
unsigned int extra_delta_offset64;
+int do_smi;
double bclk;
unsigned int show_pkg;
unsigned int show_core;
@@ -74,6 +74,43 @@ unsigned int show_cpu;
unsigned int show_pkg_only;
unsigned int show_core_only;
char *output_buffer, *outp;
+unsigned int do_rapl;
+unsigned int do_dts;
+unsigned int do_ptm;
+unsigned int tcc_activation_temp;
+unsigned int tcc_activation_temp_override;
+double rapl_power_units, rapl_energy_units, rapl_time_units;
+double rapl_joule_counter_range;
+
+#define RAPL_PKG (1 << 0)
+ /* 0x610 MSR_PKG_POWER_LIMIT */
+ /* 0x611 MSR_PKG_ENERGY_STATUS */
+#define RAPL_PKG_PERF_STATUS (1 << 1)
+ /* 0x613 MSR_PKG_PERF_STATUS */
+#define RAPL_PKG_POWER_INFO (1 << 2)
+ /* 0x614 MSR_PKG_POWER_INFO */
+
+#define RAPL_DRAM (1 << 3)
+ /* 0x618 MSR_DRAM_POWER_LIMIT */
+ /* 0x619 MSR_DRAM_ENERGY_STATUS */
+ /* 0x61c MSR_DRAM_POWER_INFO */
+#define RAPL_DRAM_PERF_STATUS (1 << 4)
+ /* 0x61b MSR_DRAM_PERF_STATUS */
+
+#define RAPL_CORES (1 << 5)
+ /* 0x638 MSR_PP0_POWER_LIMIT */
+ /* 0x639 MSR_PP0_ENERGY_STATUS */
+#define RAPL_CORE_POLICY (1 << 6)
+ /* 0x63a MSR_PP0_POLICY */
+
+
+#define RAPL_GFX (1 << 7)
+ /* 0x640 MSR_PP1_POWER_LIMIT */
+ /* 0x641 MSR_PP1_ENERGY_STATUS */
+ /* 0x642 MSR_PP1_POLICY */
+#define TJMAX_DEFAULT 100
+
+#define MAX(a, b) ((a) > (b) ? (a) : (b))
int aperf_mperf_unstable;
int backwards_count;
@@ -86,11 +123,12 @@ struct thread_data {
unsigned long long tsc;
unsigned long long aperf;
unsigned long long mperf;
- unsigned long long c1; /* derived */
+ unsigned long long c1;
unsigned long long extra_msr64;
unsigned long long extra_delta64;
unsigned long long extra_msr32;
unsigned long long extra_delta32;
+ unsigned int smi_count;
unsigned int cpu_id;
unsigned int flags;
#define CPU_IS_FIRST_THREAD_IN_CORE 0x2
@@ -101,6 +139,7 @@ struct core_data {
unsigned long long c3;
unsigned long long c6;
unsigned long long c7;
+ unsigned int core_temp_c;
unsigned int core_id;
} *core_even, *core_odd;
@@ -109,7 +148,18 @@ struct pkg_data {
unsigned long long pc3;
unsigned long long pc6;
unsigned long long pc7;
+ unsigned long long pc8;
+ unsigned long long pc9;
+ unsigned long long pc10;
unsigned int package_id;
+ unsigned int energy_pkg; /* MSR_PKG_ENERGY_STATUS */
+ unsigned int energy_dram; /* MSR_DRAM_ENERGY_STATUS */
+ unsigned int energy_cores; /* MSR_PP0_ENERGY_STATUS */
+ unsigned int energy_gfx; /* MSR_PP1_ENERGY_STATUS */
+ unsigned int rapl_pkg_perf_status; /* MSR_PKG_PERF_STATUS */
+ unsigned int rapl_dram_perf_status; /* MSR_DRAM_PERF_STATUS */
+ unsigned int pkg_temp_c;
+
} *package_even, *package_odd;
#define ODD_COUNTERS thread_odd, core_odd, package_odd
@@ -206,108 +256,171 @@ int get_msr(int cpu, off_t offset, unsigned long long *msr)
retval = pread(fd, msr, sizeof *msr, offset);
close(fd);
- if (retval != sizeof *msr)
+ if (retval != sizeof *msr) {
+ fprintf(stderr, "%s offset 0x%llx read failed\n", pathname, (unsigned long long)offset);
return -1;
+ }
return 0;
}
+/*
+ * Example Format w/ field column widths:
+ *
+ * Package Core CPU Avg_MHz Bzy_MHz TSC_MHz SMI %Busy CPU_%c1 CPU_%c3 CPU_%c6 CPU_%c7 CoreTmp PkgTmp Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt
+ * 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567
+ */
+
void print_header(void)
{
if (show_pkg)
- outp += sprintf(outp, "pk");
- if (show_pkg)
- outp += sprintf(outp, " ");
+ outp += sprintf(outp, "Package ");
if (show_core)
- outp += sprintf(outp, "cor");
+ outp += sprintf(outp, " Core ");
if (show_cpu)
- outp += sprintf(outp, " CPU");
- if (show_pkg || show_core || show_cpu)
- outp += sprintf(outp, " ");
+ outp += sprintf(outp, " CPU ");
+ if (has_aperf)
+ outp += sprintf(outp, "Avg_MHz ");
if (do_nhm_cstates)
- outp += sprintf(outp, " %%c0");
+ outp += sprintf(outp, " %%Busy ");
if (has_aperf)
- outp += sprintf(outp, " GHz");
- outp += sprintf(outp, " TSC");
+ outp += sprintf(outp, "Bzy_MHz ");
+ outp += sprintf(outp, "TSC_MHz ");
+ if (do_smi)
+ outp += sprintf(outp, " SMI ");
if (extra_delta_offset32)
- outp += sprintf(outp, " count 0x%03X", extra_delta_offset32);
+ outp += sprintf(outp, " count 0x%03X ", extra_delta_offset32);
if (extra_delta_offset64)
- outp += sprintf(outp, " COUNT 0x%03X", extra_delta_offset64);
+ outp += sprintf(outp, " COUNT 0x%03X ", extra_delta_offset64);
if (extra_msr_offset32)
- outp += sprintf(outp, " MSR 0x%03X", extra_msr_offset32);
+ outp += sprintf(outp, " MSR 0x%03X ", extra_msr_offset32);
if (extra_msr_offset64)
- outp += sprintf(outp, " MSR 0x%03X", extra_msr_offset64);
+ outp += sprintf(outp, " MSR 0x%03X ", extra_msr_offset64);
if (do_nhm_cstates)
- outp += sprintf(outp, " %%c1");
+ outp += sprintf(outp, " CPU%%c1 ");
+ if (do_nhm_cstates && !do_slm_cstates)
+ outp += sprintf(outp, " CPU%%c3 ");
if (do_nhm_cstates)
- outp += sprintf(outp, " %%c3");
- if (do_nhm_cstates)
- outp += sprintf(outp, " %%c6");
+ outp += sprintf(outp, " CPU%%c6 ");
if (do_snb_cstates)
- outp += sprintf(outp, " %%c7");
+ outp += sprintf(outp, " CPU%%c7 ");
+
+ if (do_dts)
+ outp += sprintf(outp, "CoreTmp ");
+ if (do_ptm)
+ outp += sprintf(outp, " PkgTmp ");
+
if (do_snb_cstates)
- outp += sprintf(outp, " %%pc2");
- if (do_nhm_cstates)
- outp += sprintf(outp, " %%pc3");
- if (do_nhm_cstates)
- outp += sprintf(outp, " %%pc6");
+ outp += sprintf(outp, "Pkg%%pc2 ");
+ if (do_nhm_cstates && !do_slm_cstates)
+ outp += sprintf(outp, "Pkg%%pc3 ");
+ if (do_nhm_cstates && !do_slm_cstates)
+ outp += sprintf(outp, "Pkg%%pc6 ");
if (do_snb_cstates)
- outp += sprintf(outp, " %%pc7");
+ outp += sprintf(outp, "Pkg%%pc7 ");
+ if (do_c8_c9_c10) {
+ outp += sprintf(outp, "Pkg%%pc8 ");
+ outp += sprintf(outp, "Pkg%%pc9 ");
+ outp += sprintf(outp, "Pk%%pc10 ");
+ }
+ if (do_rapl && !rapl_joules) {
+ if (do_rapl & RAPL_PKG)
+ outp += sprintf(outp, "PkgWatt ");
+ if (do_rapl & RAPL_CORES)
+ outp += sprintf(outp, "CorWatt ");
+ if (do_rapl & RAPL_GFX)
+ outp += sprintf(outp, "GFXWatt ");
+ if (do_rapl & RAPL_DRAM)
+ outp += sprintf(outp, "RAMWatt ");
+ if (do_rapl & RAPL_PKG_PERF_STATUS)
+ outp += sprintf(outp, " PKG_%% ");
+ if (do_rapl & RAPL_DRAM_PERF_STATUS)
+ outp += sprintf(outp, " RAM_%% ");
+ } else {
+ if (do_rapl & RAPL_PKG)
+ outp += sprintf(outp, " Pkg_J ");
+ if (do_rapl & RAPL_CORES)
+ outp += sprintf(outp, " Cor_J ");
+ if (do_rapl & RAPL_GFX)
+ outp += sprintf(outp, " GFX_J ");
+ if (do_rapl & RAPL_DRAM)
+ outp += sprintf(outp, " RAM_W ");
+ if (do_rapl & RAPL_PKG_PERF_STATUS)
+ outp += sprintf(outp, " PKG_%% ");
+ if (do_rapl & RAPL_DRAM_PERF_STATUS)
+ outp += sprintf(outp, " RAM_%% ");
+ outp += sprintf(outp, " time ");
+
+ }
outp += sprintf(outp, "\n");
}
int dump_counters(struct thread_data *t, struct core_data *c,
struct pkg_data *p)
{
- fprintf(stderr, "t %p, c %p, p %p\n", t, c, p);
+ outp += sprintf(outp, "t %p, c %p, p %p\n", t, c, p);
if (t) {
- fprintf(stderr, "CPU: %d flags 0x%x\n", t->cpu_id, t->flags);
- fprintf(stderr, "TSC: %016llX\n", t->tsc);
- fprintf(stderr, "aperf: %016llX\n", t->aperf);
- fprintf(stderr, "mperf: %016llX\n", t->mperf);
- fprintf(stderr, "c1: %016llX\n", t->c1);
- fprintf(stderr, "msr0x%x: %08llX\n",
+ outp += sprintf(outp, "CPU: %d flags 0x%x\n",
+ t->cpu_id, t->flags);
+ outp += sprintf(outp, "TSC: %016llX\n", t->tsc);
+ outp += sprintf(outp, "aperf: %016llX\n", t->aperf);
+ outp += sprintf(outp, "mperf: %016llX\n", t->mperf);
+ outp += sprintf(outp, "c1: %016llX\n", t->c1);
+ outp += sprintf(outp, "msr0x%x: %08llX\n",
extra_delta_offset32, t->extra_delta32);
- fprintf(stderr, "msr0x%x: %016llX\n",
+ outp += sprintf(outp, "msr0x%x: %016llX\n",
extra_delta_offset64, t->extra_delta64);
- fprintf(stderr, "msr0x%x: %08llX\n",
+ outp += sprintf(outp, "msr0x%x: %08llX\n",
extra_msr_offset32, t->extra_msr32);
- fprintf(stderr, "msr0x%x: %016llX\n",
+ outp += sprintf(outp, "msr0x%x: %016llX\n",
extra_msr_offset64, t->extra_msr64);
+ if (do_smi)
+ outp += sprintf(outp, "SMI: %08X\n", t->smi_count);
}
if (c) {
- fprintf(stderr, "core: %d\n", c->core_id);
- fprintf(stderr, "c3: %016llX\n", c->c3);
- fprintf(stderr, "c6: %016llX\n", c->c6);
- fprintf(stderr, "c7: %016llX\n", c->c7);
+ outp += sprintf(outp, "core: %d\n", c->core_id);
+ outp += sprintf(outp, "c3: %016llX\n", c->c3);
+ outp += sprintf(outp, "c6: %016llX\n", c->c6);
+ outp += sprintf(outp, "c7: %016llX\n", c->c7);
+ outp += sprintf(outp, "DTS: %dC\n", c->core_temp_c);
}
if (p) {
- fprintf(stderr, "package: %d\n", p->package_id);
- fprintf(stderr, "pc2: %016llX\n", p->pc2);
- fprintf(stderr, "pc3: %016llX\n", p->pc3);
- fprintf(stderr, "pc6: %016llX\n", p->pc6);
- fprintf(stderr, "pc7: %016llX\n", p->pc7);
+ outp += sprintf(outp, "package: %d\n", p->package_id);
+ outp += sprintf(outp, "pc2: %016llX\n", p->pc2);
+ outp += sprintf(outp, "pc3: %016llX\n", p->pc3);
+ outp += sprintf(outp, "pc6: %016llX\n", p->pc6);
+ outp += sprintf(outp, "pc7: %016llX\n", p->pc7);
+ outp += sprintf(outp, "pc8: %016llX\n", p->pc8);
+ outp += sprintf(outp, "pc9: %016llX\n", p->pc9);
+ outp += sprintf(outp, "pc10: %016llX\n", p->pc10);
+ outp += sprintf(outp, "Joules PKG: %0X\n", p->energy_pkg);
+ outp += sprintf(outp, "Joules COR: %0X\n", p->energy_cores);
+ outp += sprintf(outp, "Joules GFX: %0X\n", p->energy_gfx);
+ outp += sprintf(outp, "Joules RAM: %0X\n", p->energy_dram);
+ outp += sprintf(outp, "Throttle PKG: %0X\n",
+ p->rapl_pkg_perf_status);
+ outp += sprintf(outp, "Throttle RAM: %0X\n",
+ p->rapl_dram_perf_status);
+ outp += sprintf(outp, "PTM: %dC\n", p->pkg_temp_c);
}
+
+ outp += sprintf(outp, "\n");
+
return 0;
}
/*
* column formatting convention & formats
- * package: "pk" 2 columns %2d
- * core: "cor" 3 columns %3d
- * CPU: "CPU" 3 columns %3d
- * GHz: "GHz" 3 columns %3.2
- * TSC: "TSC" 3 columns %3.2
- * percentage " %pc3" %6.2
*/
int format_counters(struct thread_data *t, struct core_data *c,
struct pkg_data *p)
{
double interval_float;
+ char *fmt8;
/* if showing only 1st thread in core and this isn't one, bail out */
if (show_core_only && !(t->flags & CPU_IS_FIRST_THREAD_IN_CORE))
@@ -322,62 +435,52 @@ int format_counters(struct thread_data *t, struct core_data *c,
/* topo columns, print blanks on 1st (average) line */
if (t == &average.threads) {
if (show_pkg)
- outp += sprintf(outp, " ");
- if (show_pkg && show_core)
- outp += sprintf(outp, " ");
+ outp += sprintf(outp, " -");
if (show_core)
- outp += sprintf(outp, " ");
+ outp += sprintf(outp, " -");
if (show_cpu)
- outp += sprintf(outp, " " " ");
+ outp += sprintf(outp, " -");
} else {
if (show_pkg) {
if (p)
- outp += sprintf(outp, "%2d", p->package_id);
+ outp += sprintf(outp, "%8d", p->package_id);
else
- outp += sprintf(outp, " ");
+ outp += sprintf(outp, " -");
}
- if (show_pkg && show_core)
- outp += sprintf(outp, " ");
if (show_core) {
if (c)
- outp += sprintf(outp, "%3d", c->core_id);
+ outp += sprintf(outp, "%8d", c->core_id);
else
- outp += sprintf(outp, " ");
+ outp += sprintf(outp, " -");
}
if (show_cpu)
- outp += sprintf(outp, " %3d", t->cpu_id);
+ outp += sprintf(outp, "%8d", t->cpu_id);
}
+ /* AvgMHz */
+ if (has_aperf)
+ outp += sprintf(outp, "%8.0f",
+ 1.0 / units * t->aperf / interval_float);
+
/* %c0 */
if (do_nhm_cstates) {
- if (show_pkg || show_core || show_cpu)
- outp += sprintf(outp, " ");
if (!skip_c0)
- outp += sprintf(outp, "%6.2f", 100.0 * t->mperf/t->tsc);
+ outp += sprintf(outp, "%8.2f", 100.0 * t->mperf/t->tsc);
else
- outp += sprintf(outp, " ****");
+ outp += sprintf(outp, "********");
}
- /* GHz */
- if (has_aperf) {
- if (!aperf_mperf_unstable) {
- outp += sprintf(outp, " %3.2f",
- 1.0 * t->tsc / units * t->aperf /
- t->mperf / interval_float);
- } else {
- if (t->aperf > t->tsc || t->mperf > t->tsc) {
- outp += sprintf(outp, " ***");
- } else {
- outp += sprintf(outp, "%3.1f*",
- 1.0 * t->tsc /
- units * t->aperf /
- t->mperf / interval_float);
- }
- }
- }
+ /* BzyMHz */
+ if (has_aperf)
+ outp += sprintf(outp, "%8.0f",
+ 1.0 * t->tsc / units * t->aperf / t->mperf / interval_float);
/* TSC */
- outp += sprintf(outp, "%5.2f", 1.0 * t->tsc/units/interval_float);
+ outp += sprintf(outp, "%8.0f", 1.0 * t->tsc/units/interval_float);
+
+ /* SMI */
+ if (do_smi)
+ outp += sprintf(outp, "%8d", t->smi_count);
/* delta */
if (extra_delta_offset32)
@@ -396,34 +499,88 @@ int format_counters(struct thread_data *t, struct core_data *c,
if (do_nhm_cstates) {
if (!skip_c1)
- outp += sprintf(outp, " %6.2f", 100.0 * t->c1/t->tsc);
+ outp += sprintf(outp, "%8.2f", 100.0 * t->c1/t->tsc);
else
- outp += sprintf(outp, " ****");
+ outp += sprintf(outp, "********");
}
/* print per-core data only for 1st thread in core */
if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE))
goto done;
+ if (do_nhm_cstates && !do_slm_cstates)
+ outp += sprintf(outp, "%8.2f", 100.0 * c->c3/t->tsc);
if (do_nhm_cstates)
- outp += sprintf(outp, " %6.2f", 100.0 * c->c3/t->tsc);
- if (do_nhm_cstates)
- outp += sprintf(outp, " %6.2f", 100.0 * c->c6/t->tsc);
+ outp += sprintf(outp, "%8.2f", 100.0 * c->c6/t->tsc);
if (do_snb_cstates)
- outp += sprintf(outp, " %6.2f", 100.0 * c->c7/t->tsc);
+ outp += sprintf(outp, "%8.2f", 100.0 * c->c7/t->tsc);
+
+ if (do_dts)
+ outp += sprintf(outp, "%8d", c->core_temp_c);
/* print per-package data only for 1st core in package */
if (!(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE))
goto done;
+ if (do_ptm)
+ outp += sprintf(outp, "%8d", p->pkg_temp_c);
+
if (do_snb_cstates)
- outp += sprintf(outp, " %6.2f", 100.0 * p->pc2/t->tsc);
- if (do_nhm_cstates)
- outp += sprintf(outp, " %6.2f", 100.0 * p->pc3/t->tsc);
- if (do_nhm_cstates)
- outp += sprintf(outp, " %6.2f", 100.0 * p->pc6/t->tsc);
+ outp += sprintf(outp, "%8.2f", 100.0 * p->pc2/t->tsc);
+ if (do_nhm_cstates && !do_slm_cstates)
+ outp += sprintf(outp, "%8.2f", 100.0 * p->pc3/t->tsc);
+ if (do_nhm_cstates && !do_slm_cstates)
+ outp += sprintf(outp, "%8.2f", 100.0 * p->pc6/t->tsc);
if (do_snb_cstates)
- outp += sprintf(outp, " %6.2f", 100.0 * p->pc7/t->tsc);
+ outp += sprintf(outp, "%8.2f", 100.0 * p->pc7/t->tsc);
+ if (do_c8_c9_c10) {
+ outp += sprintf(outp, "%8.2f", 100.0 * p->pc8/t->tsc);
+ outp += sprintf(outp, "%8.2f", 100.0 * p->pc9/t->tsc);
+ outp += sprintf(outp, "%8.2f", 100.0 * p->pc10/t->tsc);
+ }
+
+ /*
+ * If measurement interval exceeds minimum RAPL Joule Counter range,
+ * indicate that results are suspect by printing "**" in fraction place.
+ */
+ if (interval_float < rapl_joule_counter_range)
+ fmt8 = "%8.2f";
+ else
+ fmt8 = " %6.0f**";
+
+ if (do_rapl && !rapl_joules) {
+ if (do_rapl & RAPL_PKG)
+ outp += sprintf(outp, fmt8, p->energy_pkg * rapl_energy_units / interval_float);
+ if (do_rapl & RAPL_CORES)
+ outp += sprintf(outp, fmt8, p->energy_cores * rapl_energy_units / interval_float);
+ if (do_rapl & RAPL_GFX)
+ outp += sprintf(outp, fmt8, p->energy_gfx * rapl_energy_units / interval_float);
+ if (do_rapl & RAPL_DRAM)
+ outp += sprintf(outp, fmt8, p->energy_dram * rapl_energy_units / interval_float);
+ if (do_rapl & RAPL_PKG_PERF_STATUS)
+ outp += sprintf(outp, fmt8, 100.0 * p->rapl_pkg_perf_status * rapl_time_units / interval_float);
+ if (do_rapl & RAPL_DRAM_PERF_STATUS)
+ outp += sprintf(outp, fmt8, 100.0 * p->rapl_dram_perf_status * rapl_time_units / interval_float);
+ } else {
+ if (do_rapl & RAPL_PKG)
+ outp += sprintf(outp, fmt8,
+ p->energy_pkg * rapl_energy_units);
+ if (do_rapl & RAPL_CORES)
+ outp += sprintf(outp, fmt8,
+ p->energy_cores * rapl_energy_units);
+ if (do_rapl & RAPL_GFX)
+ outp += sprintf(outp, fmt8,
+ p->energy_gfx * rapl_energy_units);
+ if (do_rapl & RAPL_DRAM)
+ outp += sprintf(outp, fmt8,
+ p->energy_dram * rapl_energy_units);
+ if (do_rapl & RAPL_PKG_PERF_STATUS)
+ outp += sprintf(outp, fmt8, 100.0 * p->rapl_pkg_perf_status * rapl_time_units / interval_float);
+ if (do_rapl & RAPL_DRAM_PERF_STATUS)
+ outp += sprintf(outp, fmt8, 100.0 * p->rapl_dram_perf_status * rapl_time_units / interval_float);
+ outp += sprintf(outp, fmt8, interval_float);
+
+ }
done:
outp += sprintf(outp, "\n");
@@ -433,6 +590,7 @@ done:
void flush_stdout()
{
fputs(output_buffer, stdout);
+ fflush(stdout);
outp = output_buffer;
}
void flush_stderr()
@@ -459,6 +617,13 @@ void format_all_counters(struct thread_data *t, struct core_data *c, struct pkg_
for_all_cpus(format_counters, t, c, p);
}
+#define DELTA_WRAP32(new, old) \
+ if (new > old) { \
+ old = new - old; \
+ } else { \
+ old = 0x100000000 + new - old; \
+ }
+
void
delta_package(struct pkg_data *new, struct pkg_data *old)
{
@@ -466,6 +631,17 @@ delta_package(struct pkg_data *new, struct pkg_data *old)
old->pc3 = new->pc3 - old->pc3;
old->pc6 = new->pc6 - old->pc6;
old->pc7 = new->pc7 - old->pc7;
+ old->pc8 = new->pc8 - old->pc8;
+ old->pc9 = new->pc9 - old->pc9;
+ old->pc10 = new->pc10 - old->pc10;
+ old->pkg_temp_c = new->pkg_temp_c;
+
+ DELTA_WRAP32(new->energy_pkg, old->energy_pkg);
+ DELTA_WRAP32(new->energy_cores, old->energy_cores);
+ DELTA_WRAP32(new->energy_gfx, old->energy_gfx);
+ DELTA_WRAP32(new->energy_dram, old->energy_dram);
+ DELTA_WRAP32(new->rapl_pkg_perf_status, old->rapl_pkg_perf_status);
+ DELTA_WRAP32(new->rapl_dram_perf_status, old->rapl_dram_perf_status);
}
void
@@ -474,6 +650,7 @@ delta_core(struct core_data *new, struct core_data *old)
old->c3 = new->c3 - old->c3;
old->c6 = new->c6 - old->c6;
old->c7 = new->c7 - old->c7;
+ old->core_temp_c = new->core_temp_c;
}
/*
@@ -486,12 +663,10 @@ delta_thread(struct thread_data *new, struct thread_data *old,
old->tsc = new->tsc - old->tsc;
/* check for TSC < 1 Mcycles over interval */
- if (old->tsc < (1000 * 1000)) {
- fprintf(stderr, "Insanely slow TSC rate, TSC stops in idle?\n");
- fprintf(stderr, "You can disable all c-states by booting with \"idle=poll\"\n");
- fprintf(stderr, "or just the deep ones with \"processor.max_cstate=1\"\n");
- exit(-3);
- }
+ if (old->tsc < (1000 * 1000))
+ errx(-3, "Insanely slow TSC rate, TSC stops in idle?\n"
+ "You can disable all c-states by booting with \"idle=poll\"\n"
+ "or just the deep ones with \"processor.max_cstate=1\"");
old->c1 = new->c1 - old->c1;
@@ -516,17 +691,24 @@ delta_thread(struct thread_data *new, struct thread_data *old,
}
- /*
- * As counter collection is not atomic,
- * it is possible for mperf's non-halted cycles + idle states
- * to exceed TSC's all cycles: show c1 = 0% in that case.
- */
- if ((old->mperf + core_delta->c3 + core_delta->c6 + core_delta->c7) > old->tsc)
- old->c1 = 0;
- else {
- /* normal case, derive c1 */
- old->c1 = old->tsc - old->mperf - core_delta->c3
+ if (use_c1_residency_msr) {
+ /*
+ * Some models have a dedicated C1 residency MSR,
+ * which should be more accurate than the derivation below.
+ */
+ } else {
+ /*
+ * As counter collection is not atomic,
+ * it is possible for mperf's non-halted cycles + idle states
+ * to exceed TSC's all cycles: show c1 = 0% in that case.
+ */
+ if ((old->mperf + core_delta->c3 + core_delta->c6 + core_delta->c7) > old->tsc)
+ old->c1 = 0;
+ else {
+ /* normal case, derive c1 */
+ old->c1 = old->tsc - old->mperf - core_delta->c3
- core_delta->c6 - core_delta->c7;
+ }
}
if (old->mperf == 0) {
@@ -544,6 +726,9 @@ delta_thread(struct thread_data *new, struct thread_data *old,
*/
old->extra_msr32 = new->extra_msr32;
old->extra_msr64 = new->extra_msr64;
+
+ if (do_smi)
+ old->smi_count = new->smi_count - old->smi_count;
}
int delta_cpu(struct thread_data *t, struct core_data *c,
@@ -571,6 +756,7 @@ void clear_counters(struct thread_data *t, struct core_data *c, struct pkg_data
t->mperf = 0;
t->c1 = 0;
+ t->smi_count = 0;
t->extra_delta32 = 0;
t->extra_delta64 = 0;
@@ -580,11 +766,23 @@ void clear_counters(struct thread_data *t, struct core_data *c, struct pkg_data
c->c3 = 0;
c->c6 = 0;
c->c7 = 0;
+ c->core_temp_c = 0;
p->pc2 = 0;
p->pc3 = 0;
p->pc6 = 0;
p->pc7 = 0;
+ p->pc8 = 0;
+ p->pc9 = 0;
+ p->pc10 = 0;
+
+ p->energy_pkg = 0;
+ p->energy_dram = 0;
+ p->energy_cores = 0;
+ p->energy_gfx = 0;
+ p->rapl_pkg_perf_status = 0;
+ p->rapl_dram_perf_status = 0;
+ p->pkg_temp_c = 0;
}
int sum_counters(struct thread_data *t, struct core_data *c,
struct pkg_data *p)
@@ -605,6 +803,8 @@ int sum_counters(struct thread_data *t, struct core_data *c,
average.cores.c6 += c->c6;
average.cores.c7 += c->c7;
+ average.cores.core_temp_c = MAX(average.cores.core_temp_c, c->core_temp_c);
+
/* sum per-pkg values only for 1st core in pkg */
if (!(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE))
return 0;
@@ -613,7 +813,19 @@ int sum_counters(struct thread_data *t, struct core_data *c,
average.packages.pc3 += p->pc3;
average.packages.pc6 += p->pc6;
average.packages.pc7 += p->pc7;
+ average.packages.pc8 += p->pc8;
+ average.packages.pc9 += p->pc9;
+ average.packages.pc10 += p->pc10;
+ average.packages.energy_pkg += p->energy_pkg;
+ average.packages.energy_dram += p->energy_dram;
+ average.packages.energy_cores += p->energy_cores;
+ average.packages.energy_gfx += p->energy_gfx;
+
+ average.packages.pkg_temp_c = MAX(average.packages.pkg_temp_c, p->pkg_temp_c);
+
+ average.packages.rapl_pkg_perf_status += p->rapl_pkg_perf_status;
+ average.packages.rapl_dram_perf_status += p->rapl_dram_perf_status;
return 0;
}
/*
@@ -645,6 +857,10 @@ void compute_average(struct thread_data *t, struct core_data *c,
average.packages.pc3 /= topo.num_packages;
average.packages.pc6 /= topo.num_packages;
average.packages.pc7 /= topo.num_packages;
+
+ average.packages.pc8 /= topo.num_packages;
+ average.packages.pc9 /= topo.num_packages;
+ average.packages.pc10 /= topo.num_packages;
}
static unsigned long long rdtsc(void)
@@ -665,23 +881,31 @@ static unsigned long long rdtsc(void)
int get_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p)
{
int cpu = t->cpu_id;
+ unsigned long long msr;
- if (cpu_migrate(cpu))
+ if (cpu_migrate(cpu)) {
+ fprintf(stderr, "Could not migrate to CPU %d\n", cpu);
return -1;
+ }
t->tsc = rdtsc(); /* we are running on local CPU of interest */
if (has_aperf) {
- if (get_msr(cpu, MSR_APERF, &t->aperf))
+ if (get_msr(cpu, MSR_IA32_APERF, &t->aperf))
return -3;
- if (get_msr(cpu, MSR_MPERF, &t->mperf))
+ if (get_msr(cpu, MSR_IA32_MPERF, &t->mperf))
return -4;
}
+ if (do_smi) {
+ if (get_msr(cpu, MSR_SMI_COUNT, &msr))
+ return -5;
+ t->smi_count = msr & 0xFFFFFFFF;
+ }
if (extra_delta_offset32) {
- if (get_msr(cpu, extra_delta_offset32, &t->extra_delta32))
+ if (get_msr(cpu, extra_delta_offset32, &msr))
return -5;
- t->extra_delta32 &= 0xFFFFFFFF;
+ t->extra_delta32 = msr & 0xFFFFFFFF;
}
if (extra_delta_offset64)
@@ -689,22 +913,30 @@ int get_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p)
return -5;
if (extra_msr_offset32) {
- if (get_msr(cpu, extra_msr_offset32, &t->extra_msr32))
+ if (get_msr(cpu, extra_msr_offset32, &msr))
return -5;
- t->extra_msr32 &= 0xFFFFFFFF;
+ t->extra_msr32 = msr & 0xFFFFFFFF;
}
if (extra_msr_offset64)
if (get_msr(cpu, extra_msr_offset64, &t->extra_msr64))
return -5;
+ if (use_c1_residency_msr) {
+ if (get_msr(cpu, MSR_CORE_C1_RES, &t->c1))
+ return -6;
+ }
+
/* collect core counters only for 1st thread in core */
if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE))
return 0;
- if (do_nhm_cstates) {
+ if (do_nhm_cstates && !do_slm_cstates) {
if (get_msr(cpu, MSR_CORE_C3_RESIDENCY, &c->c3))
return -6;
+ }
+
+ if (do_nhm_cstates) {
if (get_msr(cpu, MSR_CORE_C6_RESIDENCY, &c->c6))
return -7;
}
@@ -713,11 +945,18 @@ int get_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p)
if (get_msr(cpu, MSR_CORE_C7_RESIDENCY, &c->c7))
return -8;
+ if (do_dts) {
+ if (get_msr(cpu, MSR_IA32_THERM_STATUS, &msr))
+ return -9;
+ c->core_temp_c = tcc_activation_temp - ((msr >> 16) & 0x7F);
+ }
+
+
/* collect package counters only for 1st core in package */
if (!(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE))
return 0;
- if (do_nhm_cstates) {
+ if (do_nhm_cstates && !do_slm_cstates) {
if (get_msr(cpu, MSR_PKG_C3_RESIDENCY, &p->pc3))
return -9;
if (get_msr(cpu, MSR_PKG_C6_RESIDENCY, &p->pc6))
@@ -729,6 +968,49 @@ int get_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p)
if (get_msr(cpu, MSR_PKG_C7_RESIDENCY, &p->pc7))
return -12;
}
+ if (do_c8_c9_c10) {
+ if (get_msr(cpu, MSR_PKG_C8_RESIDENCY, &p->pc8))
+ return -13;
+ if (get_msr(cpu, MSR_PKG_C9_RESIDENCY, &p->pc9))
+ return -13;
+ if (get_msr(cpu, MSR_PKG_C10_RESIDENCY, &p->pc10))
+ return -13;
+ }
+ if (do_rapl & RAPL_PKG) {
+ if (get_msr(cpu, MSR_PKG_ENERGY_STATUS, &msr))
+ return -13;
+ p->energy_pkg = msr & 0xFFFFFFFF;
+ }
+ if (do_rapl & RAPL_CORES) {
+ if (get_msr(cpu, MSR_PP0_ENERGY_STATUS, &msr))
+ return -14;
+ p->energy_cores = msr & 0xFFFFFFFF;
+ }
+ if (do_rapl & RAPL_DRAM) {
+ if (get_msr(cpu, MSR_DRAM_ENERGY_STATUS, &msr))
+ return -15;
+ p->energy_dram = msr & 0xFFFFFFFF;
+ }
+ if (do_rapl & RAPL_GFX) {
+ if (get_msr(cpu, MSR_PP1_ENERGY_STATUS, &msr))
+ return -16;
+ p->energy_gfx = msr & 0xFFFFFFFF;
+ }
+ if (do_rapl & RAPL_PKG_PERF_STATUS) {
+ if (get_msr(cpu, MSR_PKG_PERF_STATUS, &msr))
+ return -16;
+ p->rapl_pkg_perf_status = msr & 0xFFFFFFFF;
+ }
+ if (do_rapl & RAPL_DRAM_PERF_STATUS) {
+ if (get_msr(cpu, MSR_DRAM_PERF_STATUS, &msr))
+ return -16;
+ p->rapl_dram_perf_status = msr & 0xFFFFFFFF;
+ }
+ if (do_ptm) {
+ if (get_msr(cpu, MSR_IA32_PACKAGE_THERM_STATUS, &msr))
+ return -17;
+ p->pkg_temp_c = tcc_activation_temp - ((msr >> 16) & 0x7F);
+ }
return 0;
}
@@ -740,10 +1022,9 @@ void print_verbose_header(void)
if (!do_nehalem_platform_info)
return;
- get_msr(0, MSR_NEHALEM_PLATFORM_INFO, &msr);
+ get_msr(0, MSR_NHM_PLATFORM_INFO, &msr);
- if (verbose > 1)
- fprintf(stderr, "MSR_NEHALEM_PLATFORM_INFO: 0x%llx\n", msr);
+ fprintf(stderr, "cpu0: MSR_NHM_PLATFORM_INFO: 0x%08llx\n", msr);
ratio = (msr >> 40) & 0xFF;
fprintf(stderr, "%d * %.0f = %.0f MHz max efficiency\n",
@@ -753,13 +1034,16 @@ void print_verbose_header(void)
fprintf(stderr, "%d * %.0f = %.0f MHz TSC frequency\n",
ratio, bclk, ratio * bclk);
+ get_msr(0, MSR_IA32_POWER_CTL, &msr);
+ fprintf(stderr, "cpu0: MSR_IA32_POWER_CTL: 0x%08llx (C1E auto-promotion: %sabled)\n",
+ msr, msr & 0x2 ? "EN" : "DIS");
+
if (!do_ivt_turbo_ratio_limit)
goto print_nhm_turbo_ratio_limits;
get_msr(0, MSR_IVT_TURBO_RATIO_LIMIT, &msr);
- if (verbose > 1)
- fprintf(stderr, "MSR_IVT_TURBO_RATIO_LIMIT: 0x%llx\n", msr);
+ fprintf(stderr, "cpu0: MSR_IVT_TURBO_RATIO_LIMIT: 0x%08llx\n", msr);
ratio = (msr >> 56) & 0xFF;
if (ratio)
@@ -802,14 +1086,58 @@ void print_verbose_header(void)
ratio, bclk, ratio * bclk);
print_nhm_turbo_ratio_limits:
+ get_msr(0, MSR_NHM_SNB_PKG_CST_CFG_CTL, &msr);
+
+#define SNB_C1_AUTO_UNDEMOTE (1UL << 27)
+#define SNB_C3_AUTO_UNDEMOTE (1UL << 28)
+
+ fprintf(stderr, "cpu0: MSR_NHM_SNB_PKG_CST_CFG_CTL: 0x%08llx", msr);
+
+ fprintf(stderr, " (%s%s%s%s%slocked: pkg-cstate-limit=%d: ",
+ (msr & SNB_C3_AUTO_UNDEMOTE) ? "UNdemote-C3, " : "",
+ (msr & SNB_C1_AUTO_UNDEMOTE) ? "UNdemote-C1, " : "",
+ (msr & NHM_C3_AUTO_DEMOTE) ? "demote-C3, " : "",
+ (msr & NHM_C1_AUTO_DEMOTE) ? "demote-C1, " : "",
+ (msr & (1 << 15)) ? "" : "UN",
+ (unsigned int)msr & 7);
+
+
+ switch(msr & 0x7) {
+ case 0:
+ fprintf(stderr, do_slm_cstates ? "no pkg states" : "pc0");
+ break;
+ case 1:
+ fprintf(stderr, do_slm_cstates ? "no pkg states" : do_snb_cstates ? "pc2" : "pc0");
+ break;
+ case 2:
+ fprintf(stderr, do_slm_cstates ? "invalid" : do_snb_cstates ? "pc6-noret" : "pc3");
+ break;
+ case 3:
+ fprintf(stderr, do_slm_cstates ? "invalid" : "pc6");
+ break;
+ case 4:
+ fprintf(stderr, do_slm_cstates ? "pc4" : "pc7");
+ break;
+ case 5:
+ fprintf(stderr, do_slm_cstates ? "invalid" : do_snb_cstates ? "pc7s" : "invalid");
+ break;
+ case 6:
+ fprintf(stderr, do_slm_cstates ? "pc6" : "invalid");
+ break;
+ case 7:
+ fprintf(stderr, do_slm_cstates ? "pc7" : "unlimited");
+ break;
+ default:
+ fprintf(stderr, "invalid");
+ }
+ fprintf(stderr, ")\n");
if (!do_nehalem_turbo_ratio_limit)
return;
- get_msr(0, MSR_NEHALEM_TURBO_RATIO_LIMIT, &msr);
+ get_msr(0, MSR_NHM_TURBO_RATIO_LIMIT, &msr);
- if (verbose > 1)
- fprintf(stderr, "MSR_NEHALEM_TURBO_RATIO_LIMIT: 0x%llx\n", msr);
+ fprintf(stderr, "cpu0: MSR_NHM_TURBO_RATIO_LIMIT: 0x%08llx\n", msr);
ratio = (msr >> 56) & 0xFF;
if (ratio)
@@ -884,24 +1212,43 @@ void free_all_buffers(void)
}
/*
+ * Open a file, and exit on failure
+ */
+FILE *fopen_or_die(const char *path, const char *mode)
+{
+ FILE *filep = fopen(path, "r");
+ if (!filep)
+ err(1, "%s: open failed", path);
+ return filep;
+}
+
+/*
+ * Parse a file containing a single int.
+ */
+int parse_int_file(const char *fmt, ...)
+{
+ va_list args;
+ char path[PATH_MAX];
+ FILE *filep;
+ int value;
+
+ va_start(args, fmt);
+ vsnprintf(path, sizeof(path), fmt, args);
+ va_end(args);
+ filep = fopen_or_die(path, "r");
+ if (fscanf(filep, "%d", &value) != 1)
+ err(1, "%s: failed to parse number from file", path);
+ fclose(filep);
+ return value;
+}
+
+/*
* cpu_is_first_sibling_in_core(cpu)
* return 1 if given CPU is 1st HT sibling in the core
*/
int cpu_is_first_sibling_in_core(int cpu)
{
- char path[64];
- FILE *filep;
- int first_cpu;
-
- sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/thread_siblings_list", cpu);
- filep = fopen(path, "r");
- if (filep == NULL) {
- perror(path);
- exit(1);
- }
- fscanf(filep, "%d", &first_cpu);
- fclose(filep);
- return (cpu == first_cpu);
+ return cpu == parse_int_file("/sys/devices/system/cpu/cpu%d/topology/thread_siblings_list", cpu);
}
/*
@@ -910,53 +1257,17 @@ int cpu_is_first_sibling_in_core(int cpu)
*/
int cpu_is_first_core_in_package(int cpu)
{
- char path[64];
- FILE *filep;
- int first_cpu;
-
- sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/core_siblings_list", cpu);
- filep = fopen(path, "r");
- if (filep == NULL) {
- perror(path);
- exit(1);
- }
- fscanf(filep, "%d", &first_cpu);
- fclose(filep);
- return (cpu == first_cpu);
+ return cpu == parse_int_file("/sys/devices/system/cpu/cpu%d/topology/core_siblings_list", cpu);
}
int get_physical_package_id(int cpu)
{
- char path[80];
- FILE *filep;
- int pkg;
-
- sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/physical_package_id", cpu);
- filep = fopen(path, "r");
- if (filep == NULL) {
- perror(path);
- exit(1);
- }
- fscanf(filep, "%d", &pkg);
- fclose(filep);
- return pkg;
+ return parse_int_file("/sys/devices/system/cpu/cpu%d/topology/physical_package_id", cpu);
}
int get_core_id(int cpu)
{
- char path[80];
- FILE *filep;
- int core;
-
- sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/core_id", cpu);
- filep = fopen(path, "r");
- if (filep == NULL) {
- perror(path);
- exit(1);
- }
- fscanf(filep, "%d", &core);
- fclose(filep);
- return core;
+ return parse_int_file("/sys/devices/system/cpu/cpu%d/topology/core_id", cpu);
}
int get_num_ht_siblings(int cpu)
@@ -968,11 +1279,7 @@ int get_num_ht_siblings(int cpu)
char character;
sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/thread_siblings_list", cpu);
- filep = fopen(path, "r");
- if (filep == NULL) {
- perror(path);
- exit(1);
- }
+ filep = fopen_or_die(path, "r");
/*
* file format:
* if a pair of number with a character between: 2 siblings (eg. 1-2, or 1,4)
@@ -1042,17 +1349,11 @@ int for_all_proc_cpus(int (func)(int))
int cpu_num;
int retval;
- fp = fopen(proc_stat, "r");
- if (fp == NULL) {
- perror(proc_stat);
- exit(1);
- }
+ fp = fopen_or_die(proc_stat, "r");
retval = fscanf(fp, "cpu %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d\n");
- if (retval != 0) {
- perror("/proc/stat format");
- exit(1);
- }
+ if (retval != 0)
+ err(1, "%s: failed to parse format", proc_stat);
while (1) {
retval = fscanf(fp, "cpu%u %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d\n", &cpu_num);
@@ -1098,13 +1399,22 @@ int mark_cpu_present(int cpu)
void turbostat_loop()
{
int retval;
+ int restarted = 0;
restart:
+ restarted++;
+
retval = for_all_cpus(get_counters, EVEN_COUNTERS);
- if (retval) {
+ if (retval < -1) {
+ exit(retval);
+ } else if (retval == -1) {
+ if (restarted > 1) {
+ exit(retval);
+ }
re_initialize();
goto restart;
}
+ restarted = 0;
gettimeofday(&tv_even, (struct timezone *)NULL);
while (1) {
@@ -1114,7 +1424,9 @@ restart:
}
sleep(interval_sec);
retval = for_all_cpus(get_counters, ODD_COUNTERS);
- if (retval) {
+ if (retval < -1) {
+ exit(retval);
+ } else if (retval == -1) {
re_initialize();
goto restart;
}
@@ -1126,7 +1438,9 @@ restart:
flush_stdout();
sleep(interval_sec);
retval = for_all_cpus(get_counters, EVEN_COUNTERS);
- if (retval) {
+ if (retval < -1) {
+ exit(retval);
+ } else if (retval == -1) {
re_initialize();
goto restart;
}
@@ -1143,19 +1457,15 @@ void check_dev_msr()
{
struct stat sb;
- if (stat("/dev/cpu/0/msr", &sb)) {
- fprintf(stderr, "no /dev/cpu/0/msr\n");
- fprintf(stderr, "Try \"# modprobe msr\"\n");
- exit(-5);
- }
+ if (stat("/dev/cpu/0/msr", &sb))
+ err(-5, "no /dev/cpu/0/msr\n"
+ "Try \"# modprobe msr\"");
}
void check_super_user()
{
- if (getuid() != 0) {
- fprintf(stderr, "must be root\n");
- exit(-6);
- }
+ if (getuid() != 0)
+ errx(-6, "must be root");
}
int has_nehalem_turbo_ratio_limit(unsigned int family, unsigned int model)
@@ -1176,6 +1486,15 @@ int has_nehalem_turbo_ratio_limit(unsigned int family, unsigned int model)
case 0x2D: /* SNB Xeon */
case 0x3A: /* IVB */
case 0x3E: /* IVB Xeon */
+ case 0x3C: /* HSW */
+ case 0x3F: /* HSX */
+ case 0x45: /* HSW */
+ case 0x46: /* HSW */
+ case 0x37: /* BYT */
+ case 0x4D: /* AVN */
+ case 0x3D: /* BDW */
+ case 0x4F: /* BDX */
+ case 0x56: /* BDX-DE */
return 1;
case 0x2E: /* Nehalem-EX Xeon - Beckton */
case 0x2F: /* Westmere-EX Xeon - Eagleton */
@@ -1199,6 +1518,330 @@ int has_ivt_turbo_ratio_limit(unsigned int family, unsigned int model)
}
}
+/*
+ * print_epb()
+ * Decode the ENERGY_PERF_BIAS MSR
+ */
+int print_epb(struct thread_data *t, struct core_data *c, struct pkg_data *p)
+{
+ unsigned long long msr;
+ char *epb_string;
+ int cpu;
+
+ if (!has_epb)
+ return 0;
+
+ cpu = t->cpu_id;
+
+ /* EPB is per-package */
+ if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE) || !(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE))
+ return 0;
+
+ if (cpu_migrate(cpu)) {
+ fprintf(stderr, "Could not migrate to CPU %d\n", cpu);
+ return -1;
+ }
+
+ if (get_msr(cpu, MSR_IA32_ENERGY_PERF_BIAS, &msr))
+ return 0;
+
+ switch (msr & 0x7) {
+ case ENERGY_PERF_BIAS_PERFORMANCE:
+ epb_string = "performance";
+ break;
+ case ENERGY_PERF_BIAS_NORMAL:
+ epb_string = "balanced";
+ break;
+ case ENERGY_PERF_BIAS_POWERSAVE:
+ epb_string = "powersave";
+ break;
+ default:
+ epb_string = "custom";
+ break;
+ }
+ fprintf(stderr, "cpu%d: MSR_IA32_ENERGY_PERF_BIAS: 0x%08llx (%s)\n", cpu, msr, epb_string);
+
+ return 0;
+}
+
+#define RAPL_POWER_GRANULARITY 0x7FFF /* 15 bit power granularity */
+#define RAPL_TIME_GRANULARITY 0x3F /* 6 bit time granularity */
+
+double get_tdp(model)
+{
+ unsigned long long msr;
+
+ if (do_rapl & RAPL_PKG_POWER_INFO)
+ if (!get_msr(0, MSR_PKG_POWER_INFO, &msr))
+ return ((msr >> 0) & RAPL_POWER_GRANULARITY) * rapl_power_units;
+
+ switch (model) {
+ case 0x37:
+ case 0x4D:
+ return 30.0;
+ default:
+ return 135.0;
+ }
+}
+
+
+/*
+ * rapl_probe()
+ *
+ * sets do_rapl, rapl_power_units, rapl_energy_units, rapl_time_units
+ */
+void rapl_probe(unsigned int family, unsigned int model)
+{
+ unsigned long long msr;
+ unsigned int time_unit;
+ double tdp;
+
+ if (!genuine_intel)
+ return;
+
+ if (family != 6)
+ return;
+
+ switch (model) {
+ case 0x2A:
+ case 0x3A:
+ case 0x3C: /* HSW */
+ case 0x45: /* HSW */
+ case 0x46: /* HSW */
+ case 0x3D: /* BDW */
+ do_rapl = RAPL_PKG | RAPL_CORES | RAPL_CORE_POLICY | RAPL_GFX | RAPL_PKG_POWER_INFO;
+ break;
+ case 0x3F: /* HSX */
+ case 0x4F: /* BDX */
+ case 0x56: /* BDX-DE */
+ do_rapl = RAPL_PKG | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_PKG_PERF_STATUS | RAPL_PKG_POWER_INFO;
+ break;
+ case 0x2D:
+ case 0x3E:
+ do_rapl = RAPL_PKG | RAPL_CORES | RAPL_CORE_POLICY | RAPL_DRAM | RAPL_PKG_PERF_STATUS | RAPL_DRAM_PERF_STATUS | RAPL_PKG_POWER_INFO;
+ break;
+ case 0x37: /* BYT */
+ case 0x4D: /* AVN */
+ do_rapl = RAPL_PKG | RAPL_CORES ;
+ break;
+ default:
+ return;
+ }
+
+ /* units on package 0, verify later other packages match */
+ if (get_msr(0, MSR_RAPL_POWER_UNIT, &msr))
+ return;
+
+ rapl_power_units = 1.0 / (1 << (msr & 0xF));
+ if (model == 0x37)
+ rapl_energy_units = 1.0 * (1 << (msr >> 8 & 0x1F)) / 1000000;
+ else
+ rapl_energy_units = 1.0 / (1 << (msr >> 8 & 0x1F));
+
+ time_unit = msr >> 16 & 0xF;
+ if (time_unit == 0)
+ time_unit = 0xA;
+
+ rapl_time_units = 1.0 / (1 << (time_unit));
+
+ tdp = get_tdp(model);
+
+ rapl_joule_counter_range = 0xFFFFFFFF * rapl_energy_units / tdp;
+ if (verbose)
+ fprintf(stderr, "RAPL: %.0f sec. Joule Counter Range, at %.0f Watts\n", rapl_joule_counter_range, tdp);
+
+ return;
+}
+
+int print_thermal(struct thread_data *t, struct core_data *c, struct pkg_data *p)
+{
+ unsigned long long msr;
+ unsigned int dts;
+ int cpu;
+
+ if (!(do_dts || do_ptm))
+ return 0;
+
+ cpu = t->cpu_id;
+
+ /* DTS is per-core, no need to print for each thread */
+ if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE))
+ return 0;
+
+ if (cpu_migrate(cpu)) {
+ fprintf(stderr, "Could not migrate to CPU %d\n", cpu);
+ return -1;
+ }
+
+ if (do_ptm && (t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) {
+ if (get_msr(cpu, MSR_IA32_PACKAGE_THERM_STATUS, &msr))
+ return 0;
+
+ dts = (msr >> 16) & 0x7F;
+ fprintf(stderr, "cpu%d: MSR_IA32_PACKAGE_THERM_STATUS: 0x%08llx (%d C)\n",
+ cpu, msr, tcc_activation_temp - dts);
+
+#ifdef THERM_DEBUG
+ if (get_msr(cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, &msr))
+ return 0;
+
+ dts = (msr >> 16) & 0x7F;
+ dts2 = (msr >> 8) & 0x7F;
+ fprintf(stderr, "cpu%d: MSR_IA32_PACKAGE_THERM_INTERRUPT: 0x%08llx (%d C, %d C)\n",
+ cpu, msr, tcc_activation_temp - dts, tcc_activation_temp - dts2);
+#endif
+ }
+
+
+ if (do_dts) {
+ unsigned int resolution;
+
+ if (get_msr(cpu, MSR_IA32_THERM_STATUS, &msr))
+ return 0;
+
+ dts = (msr >> 16) & 0x7F;
+ resolution = (msr >> 27) & 0xF;
+ fprintf(stderr, "cpu%d: MSR_IA32_THERM_STATUS: 0x%08llx (%d C +/- %d)\n",
+ cpu, msr, tcc_activation_temp - dts, resolution);
+
+#ifdef THERM_DEBUG
+ if (get_msr(cpu, MSR_IA32_THERM_INTERRUPT, &msr))
+ return 0;
+
+ dts = (msr >> 16) & 0x7F;
+ dts2 = (msr >> 8) & 0x7F;
+ fprintf(stderr, "cpu%d: MSR_IA32_THERM_INTERRUPT: 0x%08llx (%d C, %d C)\n",
+ cpu, msr, tcc_activation_temp - dts, tcc_activation_temp - dts2);
+#endif
+ }
+
+ return 0;
+}
+
+void print_power_limit_msr(int cpu, unsigned long long msr, char *label)
+{
+ fprintf(stderr, "cpu%d: %s: %sabled (%f Watts, %f sec, clamp %sabled)\n",
+ cpu, label,
+ ((msr >> 15) & 1) ? "EN" : "DIS",
+ ((msr >> 0) & 0x7FFF) * rapl_power_units,
+ (1.0 + (((msr >> 22) & 0x3)/4.0)) * (1 << ((msr >> 17) & 0x1F)) * rapl_time_units,
+ (((msr >> 16) & 1) ? "EN" : "DIS"));
+
+ return;
+}
+
+int print_rapl(struct thread_data *t, struct core_data *c, struct pkg_data *p)
+{
+ unsigned long long msr;
+ int cpu;
+
+ if (!do_rapl)
+ return 0;
+
+ /* RAPL counters are per package, so print only for 1st thread/package */
+ if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE) || !(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE))
+ return 0;
+
+ cpu = t->cpu_id;
+ if (cpu_migrate(cpu)) {
+ fprintf(stderr, "Could not migrate to CPU %d\n", cpu);
+ return -1;
+ }
+
+ if (get_msr(cpu, MSR_RAPL_POWER_UNIT, &msr))
+ return -1;
+
+ if (verbose) {
+ fprintf(stderr, "cpu%d: MSR_RAPL_POWER_UNIT: 0x%08llx "
+ "(%f Watts, %f Joules, %f sec.)\n", cpu, msr,
+ rapl_power_units, rapl_energy_units, rapl_time_units);
+ }
+ if (do_rapl & RAPL_PKG_POWER_INFO) {
+
+ if (get_msr(cpu, MSR_PKG_POWER_INFO, &msr))
+ return -5;
+
+
+ fprintf(stderr, "cpu%d: MSR_PKG_POWER_INFO: 0x%08llx (%.0f W TDP, RAPL %.0f - %.0f W, %f sec.)\n",
+ cpu, msr,
+ ((msr >> 0) & RAPL_POWER_GRANULARITY) * rapl_power_units,
+ ((msr >> 16) & RAPL_POWER_GRANULARITY) * rapl_power_units,
+ ((msr >> 32) & RAPL_POWER_GRANULARITY) * rapl_power_units,
+ ((msr >> 48) & RAPL_TIME_GRANULARITY) * rapl_time_units);
+
+ }
+ if (do_rapl & RAPL_PKG) {
+
+ if (get_msr(cpu, MSR_PKG_POWER_LIMIT, &msr))
+ return -9;
+
+ fprintf(stderr, "cpu%d: MSR_PKG_POWER_LIMIT: 0x%08llx (%slocked)\n",
+ cpu, msr, (msr >> 63) & 1 ? "": "UN");
+
+ print_power_limit_msr(cpu, msr, "PKG Limit #1");
+ fprintf(stderr, "cpu%d: PKG Limit #2: %sabled (%f Watts, %f* sec, clamp %sabled)\n",
+ cpu,
+ ((msr >> 47) & 1) ? "EN" : "DIS",
+ ((msr >> 32) & 0x7FFF) * rapl_power_units,
+ (1.0 + (((msr >> 54) & 0x3)/4.0)) * (1 << ((msr >> 49) & 0x1F)) * rapl_time_units,
+ ((msr >> 48) & 1) ? "EN" : "DIS");
+ }
+
+ if (do_rapl & RAPL_DRAM) {
+ if (get_msr(cpu, MSR_DRAM_POWER_INFO, &msr))
+ return -6;
+
+
+ fprintf(stderr, "cpu%d: MSR_DRAM_POWER_INFO,: 0x%08llx (%.0f W TDP, RAPL %.0f - %.0f W, %f sec.)\n",
+ cpu, msr,
+ ((msr >> 0) & RAPL_POWER_GRANULARITY) * rapl_power_units,
+ ((msr >> 16) & RAPL_POWER_GRANULARITY) * rapl_power_units,
+ ((msr >> 32) & RAPL_POWER_GRANULARITY) * rapl_power_units,
+ ((msr >> 48) & RAPL_TIME_GRANULARITY) * rapl_time_units);
+
+
+ if (get_msr(cpu, MSR_DRAM_POWER_LIMIT, &msr))
+ return -9;
+ fprintf(stderr, "cpu%d: MSR_DRAM_POWER_LIMIT: 0x%08llx (%slocked)\n",
+ cpu, msr, (msr >> 31) & 1 ? "": "UN");
+
+ print_power_limit_msr(cpu, msr, "DRAM Limit");
+ }
+ if (do_rapl & RAPL_CORE_POLICY) {
+ if (verbose) {
+ if (get_msr(cpu, MSR_PP0_POLICY, &msr))
+ return -7;
+
+ fprintf(stderr, "cpu%d: MSR_PP0_POLICY: %lld\n", cpu, msr & 0xF);
+ }
+ }
+ if (do_rapl & RAPL_CORES) {
+ if (verbose) {
+
+ if (get_msr(cpu, MSR_PP0_POWER_LIMIT, &msr))
+ return -9;
+ fprintf(stderr, "cpu%d: MSR_PP0_POWER_LIMIT: 0x%08llx (%slocked)\n",
+ cpu, msr, (msr >> 31) & 1 ? "": "UN");
+ print_power_limit_msr(cpu, msr, "Cores Limit");
+ }
+ }
+ if (do_rapl & RAPL_GFX) {
+ if (verbose) {
+ if (get_msr(cpu, MSR_PP1_POLICY, &msr))
+ return -8;
+
+ fprintf(stderr, "cpu%d: MSR_PP1_POLICY: %lld\n", cpu, msr & 0xF);
+
+ if (get_msr(cpu, MSR_PP1_POWER_LIMIT, &msr))
+ return -9;
+ fprintf(stderr, "cpu%d: MSR_PP1_POWER_LIMIT: 0x%08llx (%slocked)\n",
+ cpu, msr, (msr >> 31) & 1 ? "": "UN");
+ print_power_limit_msr(cpu, msr, "GFX Limit");
+ }
+ }
+ return 0;
+}
+
int is_snb(unsigned int family, unsigned int model)
{
@@ -1210,19 +1853,144 @@ int is_snb(unsigned int family, unsigned int model)
case 0x2D:
case 0x3A: /* IVB */
case 0x3E: /* IVB Xeon */
+ case 0x3C: /* HSW */
+ case 0x3F: /* HSW */
+ case 0x45: /* HSW */
+ case 0x46: /* HSW */
+ case 0x3D: /* BDW */
+ case 0x4F: /* BDX */
+ case 0x56: /* BDX-DE */
return 1;
}
return 0;
}
+int has_c8_c9_c10(unsigned int family, unsigned int model)
+{
+ if (!genuine_intel)
+ return 0;
+
+ switch (model) {
+ case 0x45: /* HSW */
+ case 0x3D: /* BDW */
+ return 1;
+ }
+ return 0;
+}
+
+
+int is_slm(unsigned int family, unsigned int model)
+{
+ if (!genuine_intel)
+ return 0;
+ switch (model) {
+ case 0x37: /* BYT */
+ case 0x4D: /* AVN */
+ return 1;
+ }
+ return 0;
+}
+
+#define SLM_BCLK_FREQS 5
+double slm_freq_table[SLM_BCLK_FREQS] = { 83.3, 100.0, 133.3, 116.7, 80.0};
+
+double slm_bclk(void)
+{
+ unsigned long long msr = 3;
+ unsigned int i;
+ double freq;
+
+ if (get_msr(0, MSR_FSB_FREQ, &msr))
+ fprintf(stderr, "SLM BCLK: unknown\n");
+
+ i = msr & 0xf;
+ if (i >= SLM_BCLK_FREQS) {
+ fprintf(stderr, "SLM BCLK[%d] invalid\n", i);
+ msr = 3;
+ }
+ freq = slm_freq_table[i];
+
+ fprintf(stderr, "SLM BCLK: %.1f Mhz\n", freq);
+
+ return freq;
+}
+
double discover_bclk(unsigned int family, unsigned int model)
{
if (is_snb(family, model))
return 100.00;
+ else if (is_slm(family, model))
+ return slm_bclk();
else
return 133.33;
}
+/*
+ * MSR_IA32_TEMPERATURE_TARGET indicates the temperature where
+ * the Thermal Control Circuit (TCC) activates.
+ * This is usually equal to tjMax.
+ *
+ * Older processors do not have this MSR, so there we guess,
+ * but also allow cmdline over-ride with -T.
+ *
+ * Several MSR temperature values are in units of degrees-C
+ * below this value, including the Digital Thermal Sensor (DTS),
+ * Package Thermal Management Sensor (PTM), and thermal event thresholds.
+ */
+int set_temperature_target(struct thread_data *t, struct core_data *c, struct pkg_data *p)
+{
+ unsigned long long msr;
+ unsigned int target_c_local;
+ int cpu;
+
+ /* tcc_activation_temp is used only for dts or ptm */
+ if (!(do_dts || do_ptm))
+ return 0;
+
+ /* this is a per-package concept */
+ if (!(t->flags & CPU_IS_FIRST_THREAD_IN_CORE) || !(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE))
+ return 0;
+
+ cpu = t->cpu_id;
+ if (cpu_migrate(cpu)) {
+ fprintf(stderr, "Could not migrate to CPU %d\n", cpu);
+ return -1;
+ }
+
+ if (tcc_activation_temp_override != 0) {
+ tcc_activation_temp = tcc_activation_temp_override;
+ fprintf(stderr, "cpu%d: Using cmdline TCC Target (%d C)\n",
+ cpu, tcc_activation_temp);
+ return 0;
+ }
+
+ /* Temperature Target MSR is Nehalem and newer only */
+ if (!do_nehalem_platform_info)
+ goto guess;
+
+ if (get_msr(0, MSR_IA32_TEMPERATURE_TARGET, &msr))
+ goto guess;
+
+ target_c_local = (msr >> 16) & 0xFF;
+
+ if (verbose)
+ fprintf(stderr, "cpu%d: MSR_IA32_TEMPERATURE_TARGET: 0x%08llx (%d C)\n",
+ cpu, msr, target_c_local);
+
+ if (!target_c_local)
+ goto guess;
+
+ tcc_activation_temp = target_c_local;
+
+ return 0;
+
+guess:
+ tcc_activation_temp = TJMAX_DEFAULT;
+ fprintf(stderr, "cpu%d: Guessing tjMax %d C, Please use -T to specify\n",
+ cpu, tcc_activation_temp);
+
+ return 0;
+}
void check_cpuid()
{
unsigned int eax, ebx, ecx, edx, max_level;
@@ -1230,16 +1998,16 @@ void check_cpuid()
eax = ebx = ecx = edx = 0;
- asm("cpuid" : "=a" (max_level), "=b" (ebx), "=c" (ecx), "=d" (edx) : "a" (0));
+ __get_cpuid(0, &max_level, &ebx, &ecx, &edx);
if (ebx == 0x756e6547 && edx == 0x49656e69 && ecx == 0x6c65746e)
genuine_intel = 1;
if (verbose)
- fprintf(stderr, "%.4s%.4s%.4s ",
+ fprintf(stderr, "CPUID(0): %.4s%.4s%.4s ",
(char *)&ebx, (char *)&edx, (char *)&ecx);
- asm("cpuid" : "=a" (fms), "=c" (ecx), "=d" (edx) : "a" (1) : "ebx");
+ __get_cpuid(1, &fms, &ebx, &ecx, &edx);
family = (fms >> 8) & 0xf;
model = (fms >> 4) & 0xf;
stepping = fms & 0xf;
@@ -1250,10 +2018,8 @@ void check_cpuid()
fprintf(stderr, "%d CPUID levels; family:model:stepping 0x%x:%x:%x (%d:%d:%d)\n",
max_level, family, model, stepping, family, model, stepping);
- if (!(edx & (1 << 5))) {
- fprintf(stderr, "CPUID: no MSR\n");
- exit(1);
- }
+ if (!(edx & (1 << 5)))
+ errx(1, "CPUID: no MSR");
/*
* check max extended function levels of CPUID.
@@ -1261,52 +2027,62 @@ void check_cpuid()
* This check is valid for both Intel and AMD.
*/
ebx = ecx = edx = 0;
- asm("cpuid" : "=a" (max_level), "=b" (ebx), "=c" (ecx), "=d" (edx) : "a" (0x80000000));
+ __get_cpuid(0x80000000, &max_level, &ebx, &ecx, &edx);
- if (max_level < 0x80000007) {
- fprintf(stderr, "CPUID: no invariant TSC (max_level 0x%x)\n", max_level);
- exit(1);
- }
+ if (max_level < 0x80000007)
+ errx(1, "CPUID: no invariant TSC (max_level 0x%x)", max_level);
/*
* Non-Stop TSC is advertised by CPUID.EAX=0x80000007: EDX.bit8
* this check is valid for both Intel and AMD
*/
- asm("cpuid" : "=a" (eax), "=b" (ebx), "=c" (ecx), "=d" (edx) : "a" (0x80000007));
+ __get_cpuid(0x80000007, &eax, &ebx, &ecx, &edx);
has_invariant_tsc = edx & (1 << 8);
- if (!has_invariant_tsc) {
- fprintf(stderr, "No invariant TSC\n");
- exit(1);
- }
+ if (!has_invariant_tsc)
+ errx(1, "No invariant TSC");
/*
* APERF/MPERF is advertised by CPUID.EAX=0x6: ECX.bit0
* this check is valid for both Intel and AMD
*/
- asm("cpuid" : "=a" (eax), "=b" (ebx), "=c" (ecx), "=d" (edx) : "a" (0x6));
+ __get_cpuid(0x6, &eax, &ebx, &ecx, &edx);
has_aperf = ecx & (1 << 0);
- if (!has_aperf) {
- fprintf(stderr, "No APERF MSR\n");
- exit(1);
- }
+ do_dts = eax & (1 << 0);
+ do_ptm = eax & (1 << 6);
+ has_epb = ecx & (1 << 3);
+
+ if (verbose)
+ fprintf(stderr, "CPUID(6): %s%s%s%s\n",
+ has_aperf ? "APERF" : "No APERF!",
+ do_dts ? ", DTS" : "",
+ do_ptm ? ", PTM": "",
+ has_epb ? ", EPB": "");
+
+ if (!has_aperf)
+ errx(-1, "No APERF");
do_nehalem_platform_info = genuine_intel && has_invariant_tsc;
do_nhm_cstates = genuine_intel; /* all Intel w/ non-stop TSC have NHM counters */
+ do_smi = do_nhm_cstates;
do_snb_cstates = is_snb(family, model);
+ do_c8_c9_c10 = has_c8_c9_c10(family, model);
+ do_slm_cstates = is_slm(family, model);
bclk = discover_bclk(family, model);
do_nehalem_turbo_ratio_limit = has_nehalem_turbo_ratio_limit(family, model);
do_ivt_turbo_ratio_limit = has_ivt_turbo_ratio_limit(family, model);
+ rapl_probe(family, model);
+
+ return;
}
void usage()
{
- fprintf(stderr, "%s: [-v][-p|-P|-S][-c MSR# | -s]][-C MSR#][-m MSR#][-M MSR#][-i interval_sec | command ...]\n",
- progname);
- exit(1);
+ errx(1, "%s: [-v][-R][-T][-p|-P|-S][-c MSR#][-C MSR#][-m MSR#][-M MSR#][-i interval_sec | command ...]\n",
+ progname);
}
@@ -1349,19 +2125,15 @@ void topology_probe()
fprintf(stderr, "num_cpus %d max_cpu_num %d\n", topo.num_cpus, topo.max_cpu_num);
cpus = calloc(1, (topo.max_cpu_num + 1) * sizeof(struct cpu_topology));
- if (cpus == NULL) {
- perror("calloc cpus");
- exit(1);
- }
+ if (cpus == NULL)
+ err(1, "calloc cpus");
/*
* Allocate and initialize cpu_present_set
*/
cpu_present_set = CPU_ALLOC((topo.max_cpu_num + 1));
- if (cpu_present_set == NULL) {
- perror("CPU_ALLOC");
- exit(3);
- }
+ if (cpu_present_set == NULL)
+ err(3, "CPU_ALLOC");
cpu_present_setsize = CPU_ALLOC_SIZE((topo.max_cpu_num + 1));
CPU_ZERO_S(cpu_present_setsize, cpu_present_set);
for_all_proc_cpus(mark_cpu_present);
@@ -1370,10 +2142,8 @@ void topology_probe()
* Allocate and initialize cpu_affinity_set
*/
cpu_affinity_set = CPU_ALLOC((topo.max_cpu_num + 1));
- if (cpu_affinity_set == NULL) {
- perror("CPU_ALLOC");
- exit(3);
- }
+ if (cpu_affinity_set == NULL)
+ err(3, "CPU_ALLOC");
cpu_affinity_setsize = CPU_ALLOC_SIZE((topo.max_cpu_num + 1));
CPU_ZERO_S(cpu_affinity_setsize, cpu_affinity_set);
@@ -1457,8 +2227,7 @@ allocate_counters(struct thread_data **t, struct core_data **c, struct pkg_data
return;
error:
- perror("calloc counters");
- exit(1);
+ err(1, "calloc counters");
}
/*
* init_counter()
@@ -1513,12 +2282,10 @@ int initialize_counters(int cpu_id)
void allocate_output_buffer()
{
- output_buffer = calloc(1, (1 + topo.num_cpus) * 128);
+ output_buffer = calloc(1, (1 + topo.num_cpus) * 1024);
outp = output_buffer;
- if (outp == NULL) {
- perror("calloc");
- exit(-1);
- }
+ if (outp == NULL)
+ err(-1, "calloc output buffer");
}
void setup_all_buffers(void)
@@ -1529,6 +2296,7 @@ void setup_all_buffers(void)
allocate_output_buffer();
for_all_proc_cpus(initialize_counters);
}
+
void turbostat_init()
{
check_cpuid();
@@ -1540,13 +2308,27 @@ void turbostat_init()
if (verbose)
print_verbose_header();
+
+ if (verbose)
+ for_all_cpus(print_epb, ODD_COUNTERS);
+
+ if (verbose)
+ for_all_cpus(print_rapl, ODD_COUNTERS);
+
+ for_all_cpus(set_temperature_target, ODD_COUNTERS);
+
+ if (verbose)
+ for_all_cpus(print_thermal, ODD_COUNTERS);
}
int fork_it(char **argv)
{
pid_t child_pid;
+ int status;
- for_all_cpus(get_counters, EVEN_COUNTERS);
+ status = for_all_cpus(get_counters, EVEN_COUNTERS);
+ if (status)
+ exit(status);
/* clear affinity side-effect of get_counters() */
sched_setaffinity(0, cpu_present_setsize, cpu_present_set);
gettimeofday(&tv_even, (struct timezone *)NULL);
@@ -1556,20 +2338,15 @@ int fork_it(char **argv)
/* child */
execvp(argv[0], argv);
} else {
- int status;
/* parent */
- if (child_pid == -1) {
- perror("fork");
- exit(1);
- }
+ if (child_pid == -1)
+ err(1, "fork");
signal(SIGINT, SIG_IGN);
signal(SIGQUIT, SIG_IGN);
- if (waitpid(child_pid, &status, 0) == -1) {
- perror("wait");
- exit(1);
- }
+ if (waitpid(child_pid, &status, 0) == -1)
+ err(status, "waitpid");
}
/*
* n.b. fork_it() does not check for errors from for_all_cpus()
@@ -1585,7 +2362,24 @@ int fork_it(char **argv)
fprintf(stderr, "%.6f sec\n", tv_delta.tv_sec + tv_delta.tv_usec/1000000.0);
- return 0;
+ return status;
+}
+
+int get_and_dump_counters(void)
+{
+ int status;
+
+ status = for_all_cpus(get_counters, ODD_COUNTERS);
+ if (status)
+ return status;
+
+ status = for_all_cpus(dump_counters, ODD_COUNTERS);
+ if (status)
+ return status;
+
+ flush_stdout();
+
+ return status;
}
void cmdline(int argc, char **argv)
@@ -1594,7 +2388,7 @@ void cmdline(int argc, char **argv)
progname = argv[0];
- while ((opt = getopt(argc, argv, "+pPSvisc:sC:m:M:")) != -1) {
+ while ((opt = getopt(argc, argv, "+pPsSvi:c:C:m:M:RJT:")) != -1) {
switch (opt) {
case 'p':
show_core_only++;
@@ -1602,6 +2396,9 @@ void cmdline(int argc, char **argv)
case 'P':
show_pkg_only++;
break;
+ case 's':
+ dump_only++;
+ break;
case 'S':
summary_only++;
break;
@@ -1614,9 +2411,6 @@ void cmdline(int argc, char **argv)
case 'c':
sscanf(optarg, "%x", &extra_delta_offset32);
break;
- case 's':
- extra_delta_offset32 = 0x34; /* SMI counter */
- break;
case 'C':
sscanf(optarg, "%x", &extra_delta_offset64);
break;
@@ -1626,6 +2420,16 @@ void cmdline(int argc, char **argv)
case 'M':
sscanf(optarg, "%x", &extra_msr_offset64);
break;
+ case 'R':
+ rapl_verbose++;
+ break;
+ case 'T':
+ tcc_activation_temp_override = atoi(optarg);
+ break;
+ case 'J':
+ rapl_joules++;
+ break;
+
default:
usage();
}
@@ -1636,12 +2440,16 @@ int main(int argc, char **argv)
{
cmdline(argc, argv);
- if (verbose > 1)
- fprintf(stderr, "turbostat v2.1 October 6, 2012"
+ if (verbose)
+ fprintf(stderr, "turbostat v3.7 Feb 6, 2014"
" - Len Brown <lenb@kernel.org>\n");
turbostat_init();
+ /* dump counters and exit */
+ if (dump_only)
+ return get_and_dump_counters();
+
/*
* if any params left, it must be a command to fork
*/
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-11-10 05:17:23 +0000