diff options
-rw-r--r-- | arch/x86/kernel/acpi/cstate.c | 4 | ||||
-rw-r--r-- | drivers/idle/intel_idle.c | 204 | ||||
-rw-r--r-- | tools/power/x86/turbostat/turbostat.8 | 127 | ||||
-rw-r--r-- | tools/power/x86/turbostat/turbostat.c | 240 |
4 files changed, 362 insertions, 213 deletions
diff --git a/arch/x86/kernel/acpi/cstate.c b/arch/x86/kernel/acpi/cstate.c index e69182fd01cf..4b28159e0421 100644 --- a/arch/x86/kernel/acpi/cstate.c +++ b/arch/x86/kernel/acpi/cstate.c | |||
@@ -87,7 +87,9 @@ static long acpi_processor_ffh_cstate_probe_cpu(void *_cx) | |||
87 | num_cstate_subtype = edx_part & MWAIT_SUBSTATE_MASK; | 87 | num_cstate_subtype = edx_part & MWAIT_SUBSTATE_MASK; |
88 | 88 | ||
89 | retval = 0; | 89 | retval = 0; |
90 | if (num_cstate_subtype < (cx->address & MWAIT_SUBSTATE_MASK)) { | 90 | /* If the HW does not support any sub-states in this C-state */ |
91 | if (num_cstate_subtype == 0) { | ||
92 | pr_warn(FW_BUG "ACPI MWAIT C-state 0x%x not supported by HW (0x%x)\n", cx->address, edx_part); | ||
91 | retval = -1; | 93 | retval = -1; |
92 | goto out; | 94 | goto out; |
93 | } | 95 | } |
diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c index 8e1939f564f4..9ddbf550bafa 100644 --- a/drivers/idle/intel_idle.c +++ b/drivers/idle/intel_idle.c | |||
@@ -196,6 +196,53 @@ static struct cpuidle_state snb_cstates[] = { | |||
196 | .enter = NULL } | 196 | .enter = NULL } |
197 | }; | 197 | }; |
198 | 198 | ||
199 | static struct cpuidle_state byt_cstates[] = { | ||
200 | { | ||
201 | .name = "C1-BYT", | ||
202 | .desc = "MWAIT 0x00", | ||
203 | .flags = MWAIT2flg(0x00) | CPUIDLE_FLAG_TIME_VALID, | ||
204 | .exit_latency = 1, | ||
205 | .target_residency = 1, | ||
206 | .enter = &intel_idle }, | ||
207 | { | ||
208 | .name = "C1E-BYT", | ||
209 | .desc = "MWAIT 0x01", | ||
210 | .flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_TIME_VALID, | ||
211 | .exit_latency = 15, | ||
212 | .target_residency = 30, | ||
213 | .enter = &intel_idle }, | ||
214 | { | ||
215 | .name = "C6N-BYT", | ||
216 | .desc = "MWAIT 0x58", | ||
217 | .flags = MWAIT2flg(0x58) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
218 | .exit_latency = 40, | ||
219 | .target_residency = 275, | ||
220 | .enter = &intel_idle }, | ||
221 | { | ||
222 | .name = "C6S-BYT", | ||
223 | .desc = "MWAIT 0x52", | ||
224 | .flags = MWAIT2flg(0x52) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
225 | .exit_latency = 140, | ||
226 | .target_residency = 560, | ||
227 | .enter = &intel_idle }, | ||
228 | { | ||
229 | .name = "C7-BYT", | ||
230 | .desc = "MWAIT 0x60", | ||
231 | .flags = MWAIT2flg(0x60) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
232 | .exit_latency = 1200, | ||
233 | .target_residency = 1500, | ||
234 | .enter = &intel_idle }, | ||
235 | { | ||
236 | .name = "C7S-BYT", | ||
237 | .desc = "MWAIT 0x64", | ||
238 | .flags = MWAIT2flg(0x64) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
239 | .exit_latency = 10000, | ||
240 | .target_residency = 20000, | ||
241 | .enter = &intel_idle }, | ||
242 | { | ||
243 | .enter = NULL } | ||
244 | }; | ||
245 | |||
199 | static struct cpuidle_state ivb_cstates[] = { | 246 | static struct cpuidle_state ivb_cstates[] = { |
200 | { | 247 | { |
201 | .name = "C1-IVB", | 248 | .name = "C1-IVB", |
@@ -236,6 +283,105 @@ static struct cpuidle_state ivb_cstates[] = { | |||
236 | .enter = NULL } | 283 | .enter = NULL } |
237 | }; | 284 | }; |
238 | 285 | ||
286 | static struct cpuidle_state ivt_cstates[] = { | ||
287 | { | ||
288 | .name = "C1-IVT", | ||
289 | .desc = "MWAIT 0x00", | ||
290 | .flags = MWAIT2flg(0x00) | CPUIDLE_FLAG_TIME_VALID, | ||
291 | .exit_latency = 1, | ||
292 | .target_residency = 1, | ||
293 | .enter = &intel_idle }, | ||
294 | { | ||
295 | .name = "C1E-IVT", | ||
296 | .desc = "MWAIT 0x01", | ||
297 | .flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_TIME_VALID, | ||
298 | .exit_latency = 10, | ||
299 | .target_residency = 80, | ||
300 | .enter = &intel_idle }, | ||
301 | { | ||
302 | .name = "C3-IVT", | ||
303 | .desc = "MWAIT 0x10", | ||
304 | .flags = MWAIT2flg(0x10) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
305 | .exit_latency = 59, | ||
306 | .target_residency = 156, | ||
307 | .enter = &intel_idle }, | ||
308 | { | ||
309 | .name = "C6-IVT", | ||
310 | .desc = "MWAIT 0x20", | ||
311 | .flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
312 | .exit_latency = 82, | ||
313 | .target_residency = 300, | ||
314 | .enter = &intel_idle }, | ||
315 | { | ||
316 | .enter = NULL } | ||
317 | }; | ||
318 | |||
319 | static struct cpuidle_state ivt_cstates_4s[] = { | ||
320 | { | ||
321 | .name = "C1-IVT-4S", | ||
322 | .desc = "MWAIT 0x00", | ||
323 | .flags = MWAIT2flg(0x00) | CPUIDLE_FLAG_TIME_VALID, | ||
324 | .exit_latency = 1, | ||
325 | .target_residency = 1, | ||
326 | .enter = &intel_idle }, | ||
327 | { | ||
328 | .name = "C1E-IVT-4S", | ||
329 | .desc = "MWAIT 0x01", | ||
330 | .flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_TIME_VALID, | ||
331 | .exit_latency = 10, | ||
332 | .target_residency = 250, | ||
333 | .enter = &intel_idle }, | ||
334 | { | ||
335 | .name = "C3-IVT-4S", | ||
336 | .desc = "MWAIT 0x10", | ||
337 | .flags = MWAIT2flg(0x10) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
338 | .exit_latency = 59, | ||
339 | .target_residency = 300, | ||
340 | .enter = &intel_idle }, | ||
341 | { | ||
342 | .name = "C6-IVT-4S", | ||
343 | .desc = "MWAIT 0x20", | ||
344 | .flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
345 | .exit_latency = 84, | ||
346 | .target_residency = 400, | ||
347 | .enter = &intel_idle }, | ||
348 | { | ||
349 | .enter = NULL } | ||
350 | }; | ||
351 | |||
352 | static struct cpuidle_state ivt_cstates_8s[] = { | ||
353 | { | ||
354 | .name = "C1-IVT-8S", | ||
355 | .desc = "MWAIT 0x00", | ||
356 | .flags = MWAIT2flg(0x00) | CPUIDLE_FLAG_TIME_VALID, | ||
357 | .exit_latency = 1, | ||
358 | .target_residency = 1, | ||
359 | .enter = &intel_idle }, | ||
360 | { | ||
361 | .name = "C1E-IVT-8S", | ||
362 | .desc = "MWAIT 0x01", | ||
363 | .flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_TIME_VALID, | ||
364 | .exit_latency = 10, | ||
365 | .target_residency = 500, | ||
366 | .enter = &intel_idle }, | ||
367 | { | ||
368 | .name = "C3-IVT-8S", | ||
369 | .desc = "MWAIT 0x10", | ||
370 | .flags = MWAIT2flg(0x10) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
371 | .exit_latency = 59, | ||
372 | .target_residency = 600, | ||
373 | .enter = &intel_idle }, | ||
374 | { | ||
375 | .name = "C6-IVT-8S", | ||
376 | .desc = "MWAIT 0x20", | ||
377 | .flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TIME_VALID | CPUIDLE_FLAG_TLB_FLUSHED, | ||
378 | .exit_latency = 88, | ||
379 | .target_residency = 700, | ||
380 | .enter = &intel_idle }, | ||
381 | { | ||
382 | .enter = NULL } | ||
383 | }; | ||
384 | |||
239 | static struct cpuidle_state hsw_cstates[] = { | 385 | static struct cpuidle_state hsw_cstates[] = { |
240 | { | 386 | { |
241 | .name = "C1-HSW", | 387 | .name = "C1-HSW", |
@@ -464,11 +610,21 @@ static const struct idle_cpu idle_cpu_snb = { | |||
464 | .disable_promotion_to_c1e = true, | 610 | .disable_promotion_to_c1e = true, |
465 | }; | 611 | }; |
466 | 612 | ||
613 | static const struct idle_cpu idle_cpu_byt = { | ||
614 | .state_table = byt_cstates, | ||
615 | .disable_promotion_to_c1e = true, | ||
616 | }; | ||
617 | |||
467 | static const struct idle_cpu idle_cpu_ivb = { | 618 | static const struct idle_cpu idle_cpu_ivb = { |
468 | .state_table = ivb_cstates, | 619 | .state_table = ivb_cstates, |
469 | .disable_promotion_to_c1e = true, | 620 | .disable_promotion_to_c1e = true, |
470 | }; | 621 | }; |
471 | 622 | ||
623 | static const struct idle_cpu idle_cpu_ivt = { | ||
624 | .state_table = ivt_cstates, | ||
625 | .disable_promotion_to_c1e = true, | ||
626 | }; | ||
627 | |||
472 | static const struct idle_cpu idle_cpu_hsw = { | 628 | static const struct idle_cpu idle_cpu_hsw = { |
473 | .state_table = hsw_cstates, | 629 | .state_table = hsw_cstates, |
474 | .disable_promotion_to_c1e = true, | 630 | .disable_promotion_to_c1e = true, |
@@ -494,8 +650,10 @@ static const struct x86_cpu_id intel_idle_ids[] = { | |||
494 | ICPU(0x2f, idle_cpu_nehalem), | 650 | ICPU(0x2f, idle_cpu_nehalem), |
495 | ICPU(0x2a, idle_cpu_snb), | 651 | ICPU(0x2a, idle_cpu_snb), |
496 | ICPU(0x2d, idle_cpu_snb), | 652 | ICPU(0x2d, idle_cpu_snb), |
653 | ICPU(0x36, idle_cpu_atom), | ||
654 | ICPU(0x37, idle_cpu_byt), | ||
497 | ICPU(0x3a, idle_cpu_ivb), | 655 | ICPU(0x3a, idle_cpu_ivb), |
498 | ICPU(0x3e, idle_cpu_ivb), | 656 | ICPU(0x3e, idle_cpu_ivt), |
499 | ICPU(0x3c, idle_cpu_hsw), | 657 | ICPU(0x3c, idle_cpu_hsw), |
500 | ICPU(0x3f, idle_cpu_hsw), | 658 | ICPU(0x3f, idle_cpu_hsw), |
501 | ICPU(0x45, idle_cpu_hsw), | 659 | ICPU(0x45, idle_cpu_hsw), |
@@ -572,6 +730,39 @@ static void intel_idle_cpuidle_devices_uninit(void) | |||
572 | free_percpu(intel_idle_cpuidle_devices); | 730 | free_percpu(intel_idle_cpuidle_devices); |
573 | return; | 731 | return; |
574 | } | 732 | } |
733 | |||
734 | /* | ||
735 | * intel_idle_state_table_update() | ||
736 | * | ||
737 | * Update the default state_table for this CPU-id | ||
738 | * | ||
739 | * Currently used to access tuned IVT multi-socket targets | ||
740 | * Assumption: num_sockets == (max_package_num + 1) | ||
741 | */ | ||
742 | void intel_idle_state_table_update(void) | ||
743 | { | ||
744 | /* IVT uses a different table for 1-2, 3-4, and > 4 sockets */ | ||
745 | if (boot_cpu_data.x86_model == 0x3e) { /* IVT */ | ||
746 | int cpu, package_num, num_sockets = 1; | ||
747 | |||
748 | for_each_online_cpu(cpu) { | ||
749 | package_num = topology_physical_package_id(cpu); | ||
750 | if (package_num + 1 > num_sockets) { | ||
751 | num_sockets = package_num + 1; | ||
752 | |||
753 | if (num_sockets > 4) | ||
754 | cpuidle_state_table = ivt_cstates_8s; | ||
755 | return; | ||
756 | } | ||
757 | } | ||
758 | |||
759 | if (num_sockets > 2) | ||
760 | cpuidle_state_table = ivt_cstates_4s; | ||
761 | /* else, 1 and 2 socket systems use default ivt_cstates */ | ||
762 | } | ||
763 | return; | ||
764 | } | ||
765 | |||
575 | /* | 766 | /* |
576 | * intel_idle_cpuidle_driver_init() | 767 | * intel_idle_cpuidle_driver_init() |
577 | * allocate, initialize cpuidle_states | 768 | * allocate, initialize cpuidle_states |
@@ -581,10 +772,12 @@ static int __init intel_idle_cpuidle_driver_init(void) | |||
581 | int cstate; | 772 | int cstate; |
582 | struct cpuidle_driver *drv = &intel_idle_driver; | 773 | struct cpuidle_driver *drv = &intel_idle_driver; |
583 | 774 | ||
775 | intel_idle_state_table_update(); | ||
776 | |||
584 | drv->state_count = 1; | 777 | drv->state_count = 1; |
585 | 778 | ||
586 | for (cstate = 0; cstate < CPUIDLE_STATE_MAX; ++cstate) { | 779 | for (cstate = 0; cstate < CPUIDLE_STATE_MAX; ++cstate) { |
587 | int num_substates, mwait_hint, mwait_cstate, mwait_substate; | 780 | int num_substates, mwait_hint, mwait_cstate; |
588 | 781 | ||
589 | if (cpuidle_state_table[cstate].enter == NULL) | 782 | if (cpuidle_state_table[cstate].enter == NULL) |
590 | break; | 783 | break; |
@@ -597,14 +790,13 @@ static int __init intel_idle_cpuidle_driver_init(void) | |||
597 | 790 | ||
598 | mwait_hint = flg2MWAIT(cpuidle_state_table[cstate].flags); | 791 | mwait_hint = flg2MWAIT(cpuidle_state_table[cstate].flags); |
599 | mwait_cstate = MWAIT_HINT2CSTATE(mwait_hint); | 792 | mwait_cstate = MWAIT_HINT2CSTATE(mwait_hint); |
600 | mwait_substate = MWAIT_HINT2SUBSTATE(mwait_hint); | ||
601 | 793 | ||
602 | /* does the state exist in CPUID.MWAIT? */ | 794 | /* number of sub-states for this state in CPUID.MWAIT */ |
603 | num_substates = (mwait_substates >> ((mwait_cstate + 1) * 4)) | 795 | num_substates = (mwait_substates >> ((mwait_cstate + 1) * 4)) |
604 | & MWAIT_SUBSTATE_MASK; | 796 | & MWAIT_SUBSTATE_MASK; |
605 | 797 | ||
606 | /* if sub-state in table is not enumerated by CPUID */ | 798 | /* if NO sub-states for this state in CPUID, skip it */ |
607 | if ((mwait_substate + 1) > num_substates) | 799 | if (num_substates == 0) |
608 | continue; | 800 | continue; |
609 | 801 | ||
610 | if (((mwait_cstate + 1) > 2) && | 802 | if (((mwait_cstate + 1) > 2) && |
diff --git a/tools/power/x86/turbostat/turbostat.8 b/tools/power/x86/turbostat/turbostat.8 index b4ddb748356c..56bfb523c5bb 100644 --- a/tools/power/x86/turbostat/turbostat.8 +++ b/tools/power/x86/turbostat/turbostat.8 | |||
@@ -47,21 +47,22 @@ displays the statistics gathered since it was forked. | |||
47 | .PP | 47 | .PP |
48 | .SH FIELD DESCRIPTIONS | 48 | .SH FIELD DESCRIPTIONS |
49 | .nf | 49 | .nf |
50 | \fBpk\fP processor package number. | 50 | \fBPackage\fP processor package number. |
51 | \fBcor\fP processor core number. | 51 | \fBCore\fP processor core number. |
52 | \fBCPU\fP Linux CPU (logical processor) number. | 52 | \fBCPU\fP Linux CPU (logical processor) number. |
53 | Note that multiple CPUs per core indicate support for Intel(R) Hyper-Threading Technology. | 53 | Note that multiple CPUs per core indicate support for Intel(R) Hyper-Threading Technology. |
54 | \fB%c0\fP percent of the interval that the CPU retired instructions. | 54 | \fBAVG_MHz\fP number of cycles executed divided by time elapsed. |
55 | \fBGHz\fP average clock rate while the CPU was in c0 state. | 55 | \fB%Buzy\fP percent of the interval that the CPU retired instructions, aka. % of time in "C0" state. |
56 | \fBTSC\fP average GHz that the TSC ran during the entire interval. | 56 | \fBBzy_MHz\fP average clock rate while the CPU was busy (in "c0" state). |
57 | \fB%c1, %c3, %c6, %c7\fP show the percentage residency in hardware core idle states. | 57 | \fBTSC_MHz\fP average MHz that the TSC ran during the entire interval. |
58 | \fBCTMP\fP Degrees Celsius reported by the per-core Digital Thermal Sensor. | 58 | \fBCPU%c1, CPU%c3, CPU%c6, CPU%c7\fP show the percentage residency in hardware core idle states. |
59 | \fBPTMP\fP Degrees Celsius reported by the per-package Package Thermal Monitor. | 59 | \fBCoreTmp\fP Degrees Celsius reported by the per-core Digital Thermal Sensor. |
60 | \fB%pc2, %pc3, %pc6, %pc7\fP percentage residency in hardware package idle states. | 60 | \fBPkgTtmp\fP Degrees Celsius reported by the per-package Package Thermal Monitor. |
61 | \fBPkg_W\fP Watts consumed by the whole package. | 61 | \fBPkg%pc2, Pkg%pc3, Pkg%pc6, Pkg%pc7\fP percentage residency in hardware package idle states. |
62 | \fBCor_W\fP Watts consumed by the core part of the package. | 62 | \fBPkgWatt\fP Watts consumed by the whole package. |
63 | \fBGFX_W\fP Watts consumed by the Graphics part of the package -- available only on client processors. | 63 | \fBCorWatt\fP Watts consumed by the core part of the package. |
64 | \fBRAM_W\fP Watts consumed by the DRAM DIMMS -- available only on server processors. | 64 | \fBGFXWatt\fP Watts consumed by the Graphics part of the package -- available only on client processors. |
65 | \fBRAMWatt\fP Watts consumed by the DRAM DIMMS -- available only on server processors. | ||
65 | \fBPKG_%\fP percent of the interval that RAPL throttling was active on the Package. | 66 | \fBPKG_%\fP percent of the interval that RAPL throttling was active on the Package. |
66 | \fBRAM_%\fP percent of the interval that RAPL throttling was active on DRAM. | 67 | \fBRAM_%\fP percent of the interval that RAPL throttling was active on DRAM. |
67 | .fi | 68 | .fi |
@@ -78,29 +79,17 @@ For Watts columns, the summary is a system total. | |||
78 | Subsequent rows show per-CPU statistics. | 79 | Subsequent rows show per-CPU statistics. |
79 | 80 | ||
80 | .nf | 81 | .nf |
81 | [root@sandy]# ./turbostat | 82 | [root@ivy]# ./turbostat |
82 | cor CPU %c0 GHz TSC %c1 %c3 %c6 %c7 CTMP PTMP %pc2 %pc3 %pc6 %pc7 Pkg_W Cor_W GFX_W | 83 | Core CPU Avg_MHz %Busy Bzy_MHz TSC_MHz SMI CPU%c1 CPU%c3 CPU%c6 CPU%c7 CoreTmp PkgTmp Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt |
83 | 0.06 0.80 2.29 0.11 0.00 0.00 99.83 47 40 0.26 0.01 0.44 98.78 3.49 0.12 0.14 | 84 | - - 6 0.36 1596 3492 0 0.59 0.01 99.04 0.00 23 24 23.82 0.01 72.47 0.00 6.40 1.01 0.00 |
84 | 0 0 0.07 0.80 2.29 0.07 0.00 0.00 99.86 40 40 0.26 0.01 0.44 98.78 3.49 0.12 0.14 | 85 | 0 0 9 0.58 1596 3492 0 0.28 0.01 99.13 0.00 23 24 23.82 0.01 72.47 0.00 6.40 1.01 0.00 |
85 | 0 4 0.03 0.80 2.29 0.12 | 86 | 0 4 1 0.07 1596 3492 0 0.79 |
86 | 1 1 0.04 0.80 2.29 0.25 0.01 0.00 99.71 40 | 87 | 1 1 10 0.65 1596 3492 0 0.59 0.00 98.76 0.00 23 |
87 | 1 5 0.16 0.80 2.29 0.13 | 88 | 1 5 5 0.28 1596 3492 0 0.95 |
88 | 2 2 0.05 0.80 2.29 0.06 0.01 0.00 99.88 40 | 89 | 2 2 10 0.66 1596 3492 0 0.41 0.01 98.92 0.00 23 |
89 | 2 6 0.03 0.80 2.29 0.08 | 90 | 2 6 2 0.10 1597 3492 0 0.97 |
90 | 3 3 0.05 0.80 2.29 0.08 0.00 0.00 99.87 47 | 91 | 3 3 3 0.20 1596 3492 0 0.44 0.00 99.37 0.00 23 |
91 | 3 7 0.04 0.84 2.29 0.09 | 92 | 3 7 5 0.31 1596 3492 0 0.33 |
92 | .fi | ||
93 | .SH SUMMARY EXAMPLE | ||
94 | The "-s" option prints the column headers just once, | ||
95 | and then the one line system summary for each sample interval. | ||
96 | |||
97 | .nf | ||
98 | [root@wsm]# turbostat -S | ||
99 | %c0 GHz TSC %c1 %c3 %c6 CTMP %pc3 %pc6 | ||
100 | 1.40 2.81 3.38 10.78 43.47 44.35 42 13.67 2.09 | ||
101 | 1.34 2.90 3.38 11.48 58.96 28.23 41 19.89 0.15 | ||
102 | 1.55 2.72 3.38 26.73 37.66 34.07 42 2.53 2.80 | ||
103 | 1.37 2.83 3.38 16.95 60.05 21.63 42 5.76 0.20 | ||
104 | .fi | 93 | .fi |
105 | .SH VERBOSE EXAMPLE | 94 | .SH VERBOSE EXAMPLE |
106 | The "-v" option adds verbosity to the output: | 95 | The "-v" option adds verbosity to the output: |
@@ -154,55 +143,35 @@ eg. Here a cycle soaker is run on 1 CPU (see %c0) for a few seconds | |||
154 | until ^C while the other CPUs are mostly idle: | 143 | until ^C while the other CPUs are mostly idle: |
155 | 144 | ||
156 | .nf | 145 | .nf |
157 | [root@x980 lenb]# ./turbostat cat /dev/zero > /dev/null | 146 | root@ivy: turbostat cat /dev/zero > /dev/null |
158 | ^C | 147 | ^C |
159 | cor CPU %c0 GHz TSC %c1 %c3 %c6 %pc3 %pc6 | 148 | Core CPU Avg_MHz %Busy Bzy_MHz TSC_MHz SMI CPU%c1 CPU%c3 CPU%c6 CPU%c7 CoreTmp PkgTmp Pkg%pc2 Pkg%pc3 Pkg%pc6 Pkg%pc7 PkgWatt CorWatt GFXWatt |
160 | 8.86 3.61 3.38 15.06 31.19 44.89 0.00 0.00 | 149 | - - 496 12.75 3886 3492 0 13.16 0.04 74.04 0.00 36 36 0.00 0.00 0.00 0.00 23.15 17.65 0.00 |
161 | 0 0 1.46 3.22 3.38 16.84 29.48 52.22 0.00 0.00 | 150 | 0 0 22 0.57 3830 3492 0 0.83 0.02 98.59 0.00 27 36 0.00 0.00 0.00 0.00 23.15 17.65 0.00 |
162 | 0 6 0.21 3.06 3.38 18.09 | 151 | 0 4 9 0.24 3829 3492 0 1.15 |
163 | 1 2 0.53 3.33 3.38 2.80 46.40 50.27 | 152 | 1 1 4 0.09 3783 3492 0 99.91 0.00 0.00 0.00 36 |
164 | 1 8 0.89 3.47 3.38 2.44 | 153 | 1 5 3880 99.82 3888 3492 0 0.18 |
165 | 2 4 1.36 3.43 3.38 9.04 23.71 65.89 | 154 | 2 2 17 0.44 3813 3492 0 0.77 0.04 98.75 0.00 28 |
166 | 2 10 0.18 2.86 3.38 10.22 | 155 | 2 6 12 0.32 3823 3492 0 0.89 |
167 | 8 1 0.04 2.87 3.38 99.96 0.01 0.00 | 156 | 3 3 16 0.43 3844 3492 0 0.63 0.11 98.84 0.00 30 |
168 | 8 7 99.72 3.63 3.38 0.27 | 157 | 3 7 4 0.11 3827 3492 0 0.94 |
169 | 9 3 0.31 3.21 3.38 7.64 56.55 35.50 | 158 | 30.372243 sec |
170 | 9 9 0.08 2.95 3.38 7.88 | 159 | |
171 | 10 5 1.42 3.43 3.38 2.14 30.99 65.44 | ||
172 | 10 11 0.16 2.88 3.38 3.40 | ||
173 | .fi | 160 | .fi |
174 | Above the cycle soaker drives cpu7 up its 3.6 GHz turbo limit | 161 | Above the cycle soaker drives cpu5 up its 3.8 GHz turbo limit |
175 | while the other processors are generally in various states of idle. | 162 | while the other processors are generally in various states of idle. |
176 | 163 | ||
177 | Note that cpu1 and cpu7 are HT siblings within core8. | 164 | Note that cpu1 and cpu5 are HT siblings within core1. |
178 | As cpu7 is very busy, it prevents its sibling, cpu1, | 165 | As cpu5 is very busy, it prevents its sibling, cpu1, |
179 | from entering a c-state deeper than c1. | 166 | from entering a c-state deeper than c1. |
180 | 167 | ||
181 | Note that turbostat reports average GHz of 3.63, while | 168 | Note that the Avg_MHz column reflects the total number of cycles executed |
182 | the arithmetic average of the GHz column above is lower. | 169 | divided by the measurement interval. If the %Busy column is 100%, |
183 | This is a weighted average, where the weight is %c0. ie. it is the total number of | 170 | then the processor was running at that speed the entire interval. |
184 | un-halted cycles elapsed per time divided by the number of CPUs. | 171 | The Avg_MHz multiplied by the %Busy results in the Bzy_MHz -- |
185 | .SH SMI COUNTING EXAMPLE | 172 | which is the average frequency while the processor was executing -- |
186 | On Intel Nehalem and newer processors, MSR 0x34 is a System Management Mode Interrupt (SMI) counter. | 173 | not including any non-busy idle time. |
187 | This counter is shown by default under the "SMI" column. | 174 | |
188 | .nf | ||
189 | [root@x980 ~]# turbostat | ||
190 | cor CPU %c0 GHz TSC SMI %c1 %c3 %c6 CTMP %pc3 %pc6 | ||
191 | 0.11 1.91 3.38 0 1.84 0.26 97.79 29 0.82 83.87 | ||
192 | 0 0 0.40 1.63 3.38 0 10.27 0.12 89.20 20 0.82 83.88 | ||
193 | 0 6 0.06 1.63 3.38 0 10.61 | ||
194 | 1 2 0.37 2.63 3.38 0 0.02 0.10 99.51 22 | ||
195 | 1 8 0.01 1.62 3.38 0 0.39 | ||
196 | 2 4 0.07 1.62 3.38 0 0.04 0.07 99.82 23 | ||
197 | 2 10 0.02 1.62 3.38 0 0.09 | ||
198 | 8 1 0.23 1.64 3.38 0 0.10 1.07 98.60 24 | ||
199 | 8 7 0.02 1.64 3.38 0 0.31 | ||
200 | 9 3 0.03 1.62 3.38 0 0.03 0.05 99.89 29 | ||
201 | 9 9 0.02 1.62 3.38 0 0.05 | ||
202 | 10 5 0.07 1.62 3.38 0 0.08 0.12 99.73 27 | ||
203 | 10 11 0.03 1.62 3.38 0 0.13 | ||
204 | ^C | ||
205 | .fi | ||
206 | .SH NOTES | 175 | .SH NOTES |
207 | 176 | ||
208 | .B "turbostat " | 177 | .B "turbostat " |
diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index 77eb130168da..7c9d8e71eb9e 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c | |||
@@ -56,7 +56,7 @@ unsigned int do_slm_cstates; | |||
56 | unsigned int use_c1_residency_msr; | 56 | unsigned int use_c1_residency_msr; |
57 | unsigned int has_aperf; | 57 | unsigned int has_aperf; |
58 | unsigned int has_epb; | 58 | unsigned int has_epb; |
59 | unsigned int units = 1000000000; /* Ghz etc */ | 59 | unsigned int units = 1000000; /* MHz etc */ |
60 | unsigned int genuine_intel; | 60 | unsigned int genuine_intel; |
61 | unsigned int has_invariant_tsc; | 61 | unsigned int has_invariant_tsc; |
62 | unsigned int do_nehalem_platform_info; | 62 | unsigned int do_nehalem_platform_info; |
@@ -264,88 +264,93 @@ int get_msr(int cpu, off_t offset, unsigned long long *msr) | |||
264 | return 0; | 264 | return 0; |
265 | } | 265 | } |
266 | 266 | ||
267 | /* | ||
268 | * Example Format w/ field column widths: | ||
269 | * | ||
270 | * 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 | ||
271 | * 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 1234567 | ||
272 | */ | ||
273 | |||
267 | void print_header(void) | 274 | void print_header(void) |
268 | { | 275 | { |
269 | if (show_pkg) | 276 | if (show_pkg) |
270 | outp += sprintf(outp, "pk"); | 277 | outp += sprintf(outp, "Package "); |
271 | if (show_pkg) | ||
272 | outp += sprintf(outp, " "); | ||
273 | if (show_core) | 278 | if (show_core) |
274 | outp += sprintf(outp, "cor"); | 279 | outp += sprintf(outp, " Core "); |
275 | if (show_cpu) | 280 | if (show_cpu) |
276 | outp += sprintf(outp, " CPU"); | 281 | outp += sprintf(outp, " CPU "); |
277 | if (show_pkg || show_core || show_cpu) | 282 | if (has_aperf) |
278 | outp += sprintf(outp, " "); | 283 | outp += sprintf(outp, "Avg_MHz "); |
279 | if (do_nhm_cstates) | 284 | if (do_nhm_cstates) |
280 | outp += sprintf(outp, " %%c0"); | 285 | outp += sprintf(outp, " %%Busy "); |
281 | if (has_aperf) | 286 | if (has_aperf) |
282 | outp += sprintf(outp, " GHz"); | 287 | outp += sprintf(outp, "Bzy_MHz "); |
283 | outp += sprintf(outp, " TSC"); | 288 | outp += sprintf(outp, "TSC_MHz "); |
284 | if (do_smi) | 289 | if (do_smi) |
285 | outp += sprintf(outp, " SMI"); | 290 | outp += sprintf(outp, " SMI "); |
286 | if (extra_delta_offset32) | 291 | if (extra_delta_offset32) |
287 | outp += sprintf(outp, " count 0x%03X", extra_delta_offset32); | 292 | outp += sprintf(outp, " count 0x%03X ", extra_delta_offset32); |
288 | if (extra_delta_offset64) | 293 | if (extra_delta_offset64) |
289 | outp += sprintf(outp, " COUNT 0x%03X", extra_delta_offset64); | 294 | outp += sprintf(outp, " COUNT 0x%03X ", extra_delta_offset64); |
290 | if (extra_msr_offset32) | 295 | if (extra_msr_offset32) |
291 | outp += sprintf(outp, " MSR 0x%03X", extra_msr_offset32); | 296 | outp += sprintf(outp, " MSR 0x%03X ", extra_msr_offset32); |
292 | if (extra_msr_offset64) | 297 | if (extra_msr_offset64) |
293 | outp += sprintf(outp, " MSR 0x%03X", extra_msr_offset64); | 298 | outp += sprintf(outp, " MSR 0x%03X ", extra_msr_offset64); |
294 | if (do_nhm_cstates) | 299 | if (do_nhm_cstates) |
295 | outp += sprintf(outp, " %%c1"); | 300 | outp += sprintf(outp, " CPU%%c1 "); |
296 | if (do_nhm_cstates && !do_slm_cstates) | 301 | if (do_nhm_cstates && !do_slm_cstates) |
297 | outp += sprintf(outp, " %%c3"); | 302 | outp += sprintf(outp, " CPU%%c3 "); |
298 | if (do_nhm_cstates) | 303 | if (do_nhm_cstates) |
299 | outp += sprintf(outp, " %%c6"); | 304 | outp += sprintf(outp, " CPU%%c6 "); |
300 | if (do_snb_cstates) | 305 | if (do_snb_cstates) |
301 | outp += sprintf(outp, " %%c7"); | 306 | outp += sprintf(outp, " CPU%%c7 "); |
302 | 307 | ||
303 | if (do_dts) | 308 | if (do_dts) |
304 | outp += sprintf(outp, " CTMP"); | 309 | outp += sprintf(outp, "CoreTmp "); |
305 | if (do_ptm) | 310 | if (do_ptm) |
306 | outp += sprintf(outp, " PTMP"); | 311 | outp += sprintf(outp, " PkgTmp "); |
307 | 312 | ||
308 | if (do_snb_cstates) | 313 | if (do_snb_cstates) |
309 | outp += sprintf(outp, " %%pc2"); | 314 | outp += sprintf(outp, "Pkg%%pc2 "); |
310 | if (do_nhm_cstates && !do_slm_cstates) | 315 | if (do_nhm_cstates && !do_slm_cstates) |
311 | outp += sprintf(outp, " %%pc3"); | 316 | outp += sprintf(outp, "Pkg%%pc3 "); |
312 | if (do_nhm_cstates && !do_slm_cstates) | 317 | if (do_nhm_cstates && !do_slm_cstates) |
313 | outp += sprintf(outp, " %%pc6"); | 318 | outp += sprintf(outp, "Pkg%%pc6 "); |
314 | if (do_snb_cstates) | 319 | if (do_snb_cstates) |
315 | outp += sprintf(outp, " %%pc7"); | 320 | outp += sprintf(outp, "Pkg%%pc7 "); |
316 | if (do_c8_c9_c10) { | 321 | if (do_c8_c9_c10) { |
317 | outp += sprintf(outp, " %%pc8"); | 322 | outp += sprintf(outp, "Pkg%%pc8 "); |
318 | outp += sprintf(outp, " %%pc9"); | 323 | outp += sprintf(outp, "Pkg%%pc9 "); |
319 | outp += sprintf(outp, " %%pc10"); | 324 | outp += sprintf(outp, "Pk%%pc10 "); |
320 | } | 325 | } |
321 | 326 | ||
322 | if (do_rapl && !rapl_joules) { | 327 | if (do_rapl && !rapl_joules) { |
323 | if (do_rapl & RAPL_PKG) | 328 | if (do_rapl & RAPL_PKG) |
324 | outp += sprintf(outp, " Pkg_W"); | 329 | outp += sprintf(outp, "PkgWatt "); |
325 | if (do_rapl & RAPL_CORES) | 330 | if (do_rapl & RAPL_CORES) |
326 | outp += sprintf(outp, " Cor_W"); | 331 | outp += sprintf(outp, "CorWatt "); |
327 | if (do_rapl & RAPL_GFX) | 332 | if (do_rapl & RAPL_GFX) |
328 | outp += sprintf(outp, " GFX_W"); | 333 | outp += sprintf(outp, "GFXWatt "); |
329 | if (do_rapl & RAPL_DRAM) | 334 | if (do_rapl & RAPL_DRAM) |
330 | outp += sprintf(outp, " RAM_W"); | 335 | outp += sprintf(outp, "RAMWatt "); |
331 | if (do_rapl & RAPL_PKG_PERF_STATUS) | 336 | if (do_rapl & RAPL_PKG_PERF_STATUS) |
332 | outp += sprintf(outp, " PKG_%%"); | 337 | outp += sprintf(outp, " PKG_%% "); |
333 | if (do_rapl & RAPL_DRAM_PERF_STATUS) | 338 | if (do_rapl & RAPL_DRAM_PERF_STATUS) |
334 | outp += sprintf(outp, " RAM_%%"); | 339 | outp += sprintf(outp, " RAM_%% "); |
335 | } else { | 340 | } else { |
336 | if (do_rapl & RAPL_PKG) | 341 | if (do_rapl & RAPL_PKG) |
337 | outp += sprintf(outp, " Pkg_J"); | 342 | outp += sprintf(outp, " Pkg_J "); |
338 | if (do_rapl & RAPL_CORES) | 343 | if (do_rapl & RAPL_CORES) |
339 | outp += sprintf(outp, " Cor_J"); | 344 | outp += sprintf(outp, " Cor_J "); |
340 | if (do_rapl & RAPL_GFX) | 345 | if (do_rapl & RAPL_GFX) |
341 | outp += sprintf(outp, " GFX_J"); | 346 | outp += sprintf(outp, " GFX_J "); |
342 | if (do_rapl & RAPL_DRAM) | 347 | if (do_rapl & RAPL_DRAM) |
343 | outp += sprintf(outp, " RAM_W"); | 348 | outp += sprintf(outp, " RAM_W "); |
344 | if (do_rapl & RAPL_PKG_PERF_STATUS) | 349 | if (do_rapl & RAPL_PKG_PERF_STATUS) |
345 | outp += sprintf(outp, " PKG_%%"); | 350 | outp += sprintf(outp, " PKG_%% "); |
346 | if (do_rapl & RAPL_DRAM_PERF_STATUS) | 351 | if (do_rapl & RAPL_DRAM_PERF_STATUS) |
347 | outp += sprintf(outp, " RAM_%%"); | 352 | outp += sprintf(outp, " RAM_%% "); |
348 | outp += sprintf(outp, " time"); | 353 | outp += sprintf(outp, " time "); |
349 | 354 | ||
350 | } | 355 | } |
351 | outp += sprintf(outp, "\n"); | 356 | outp += sprintf(outp, "\n"); |
@@ -410,25 +415,12 @@ int dump_counters(struct thread_data *t, struct core_data *c, | |||
410 | 415 | ||
411 | /* | 416 | /* |
412 | * column formatting convention & formats | 417 | * column formatting convention & formats |
413 | * package: "pk" 2 columns %2d | ||
414 | * core: "cor" 3 columns %3d | ||
415 | * CPU: "CPU" 3 columns %3d | ||
416 | * Pkg_W: %6.2 | ||
417 | * Cor_W: %6.2 | ||
418 | * GFX_W: %5.2 | ||
419 | * RAM_W: %5.2 | ||
420 | * GHz: "GHz" 3 columns %3.2 | ||
421 | * TSC: "TSC" 3 columns %3.2 | ||
422 | * SMI: "SMI" 4 columns %4d | ||
423 | * percentage " %pc3" %6.2 | ||
424 | * Perf Status percentage: %5.2 | ||
425 | * "CTMP" 4 columns %4d | ||
426 | */ | 418 | */ |
427 | int format_counters(struct thread_data *t, struct core_data *c, | 419 | int format_counters(struct thread_data *t, struct core_data *c, |
428 | struct pkg_data *p) | 420 | struct pkg_data *p) |
429 | { | 421 | { |
430 | double interval_float; | 422 | double interval_float; |
431 | char *fmt5, *fmt6; | 423 | char *fmt8; |
432 | 424 | ||
433 | /* if showing only 1st thread in core and this isn't one, bail out */ | 425 | /* if showing only 1st thread in core and this isn't one, bail out */ |
434 | if (show_core_only && !(t->flags & CPU_IS_FIRST_THREAD_IN_CORE)) | 426 | if (show_core_only && !(t->flags & CPU_IS_FIRST_THREAD_IN_CORE)) |
@@ -443,65 +435,52 @@ int format_counters(struct thread_data *t, struct core_data *c, | |||
443 | /* topo columns, print blanks on 1st (average) line */ | 435 | /* topo columns, print blanks on 1st (average) line */ |
444 | if (t == &average.threads) { | 436 | if (t == &average.threads) { |
445 | if (show_pkg) | 437 | if (show_pkg) |
446 | outp += sprintf(outp, " "); | 438 | outp += sprintf(outp, " -"); |
447 | if (show_pkg && show_core) | ||
448 | outp += sprintf(outp, " "); | ||
449 | if (show_core) | 439 | if (show_core) |
450 | outp += sprintf(outp, " "); | 440 | outp += sprintf(outp, " -"); |
451 | if (show_cpu) | 441 | if (show_cpu) |
452 | outp += sprintf(outp, " " " "); | 442 | outp += sprintf(outp, " -"); |
453 | } else { | 443 | } else { |
454 | if (show_pkg) { | 444 | if (show_pkg) { |
455 | if (p) | 445 | if (p) |
456 | outp += sprintf(outp, "%2d", p->package_id); | 446 | outp += sprintf(outp, "%8d", p->package_id); |
457 | else | 447 | else |
458 | outp += sprintf(outp, " "); | 448 | outp += sprintf(outp, " -"); |
459 | } | 449 | } |
460 | if (show_pkg && show_core) | ||
461 | outp += sprintf(outp, " "); | ||
462 | if (show_core) { | 450 | if (show_core) { |
463 | if (c) | 451 | if (c) |
464 | outp += sprintf(outp, "%3d", c->core_id); | 452 | outp += sprintf(outp, "%8d", c->core_id); |
465 | else | 453 | else |
466 | outp += sprintf(outp, " "); | 454 | outp += sprintf(outp, " -"); |
467 | } | 455 | } |
468 | if (show_cpu) | 456 | if (show_cpu) |
469 | outp += sprintf(outp, " %3d", t->cpu_id); | 457 | outp += sprintf(outp, "%8d", t->cpu_id); |
470 | } | 458 | } |
459 | |||
460 | /* AvgMHz */ | ||
461 | if (has_aperf) | ||
462 | outp += sprintf(outp, "%8.0f", | ||
463 | 1.0 / units * t->aperf / interval_float); | ||
464 | |||
471 | /* %c0 */ | 465 | /* %c0 */ |
472 | if (do_nhm_cstates) { | 466 | if (do_nhm_cstates) { |
473 | if (show_pkg || show_core || show_cpu) | ||
474 | outp += sprintf(outp, " "); | ||
475 | if (!skip_c0) | 467 | if (!skip_c0) |
476 | outp += sprintf(outp, "%6.2f", 100.0 * t->mperf/t->tsc); | 468 | outp += sprintf(outp, "%8.2f", 100.0 * t->mperf/t->tsc); |
477 | else | 469 | else |
478 | outp += sprintf(outp, " ****"); | 470 | outp += sprintf(outp, "********"); |
479 | } | 471 | } |
480 | 472 | ||
481 | /* GHz */ | 473 | /* BzyMHz */ |
482 | if (has_aperf) { | 474 | if (has_aperf) |
483 | if (!aperf_mperf_unstable) { | 475 | outp += sprintf(outp, "%8.0f", |
484 | outp += sprintf(outp, " %3.2f", | 476 | 1.0 * t->tsc / units * t->aperf / t->mperf / interval_float); |
485 | 1.0 * t->tsc / units * t->aperf / | ||
486 | t->mperf / interval_float); | ||
487 | } else { | ||
488 | if (t->aperf > t->tsc || t->mperf > t->tsc) { | ||
489 | outp += sprintf(outp, " ***"); | ||
490 | } else { | ||
491 | outp += sprintf(outp, "%3.1f*", | ||
492 | 1.0 * t->tsc / | ||
493 | units * t->aperf / | ||
494 | t->mperf / interval_float); | ||
495 | } | ||
496 | } | ||
497 | } | ||
498 | 477 | ||
499 | /* TSC */ | 478 | /* TSC */ |
500 | outp += sprintf(outp, "%5.2f", 1.0 * t->tsc/units/interval_float); | 479 | outp += sprintf(outp, "%8.0f", 1.0 * t->tsc/units/interval_float); |
501 | 480 | ||
502 | /* SMI */ | 481 | /* SMI */ |
503 | if (do_smi) | 482 | if (do_smi) |
504 | outp += sprintf(outp, "%4d", t->smi_count); | 483 | outp += sprintf(outp, "%8d", t->smi_count); |
505 | 484 | ||
506 | /* delta */ | 485 | /* delta */ |
507 | if (extra_delta_offset32) | 486 | if (extra_delta_offset32) |
@@ -520,9 +499,9 @@ int format_counters(struct thread_data *t, struct core_data *c, | |||
520 | 499 | ||
521 | if (do_nhm_cstates) { | 500 | if (do_nhm_cstates) { |
522 | if (!skip_c1) | 501 | if (!skip_c1) |
523 | outp += sprintf(outp, " %6.2f", 100.0 * t->c1/t->tsc); | 502 | outp += sprintf(outp, "%8.2f", 100.0 * t->c1/t->tsc); |
524 | else | 503 | else |
525 | outp += sprintf(outp, " ****"); | 504 | outp += sprintf(outp, "********"); |
526 | } | 505 | } |
527 | 506 | ||
528 | /* print per-core data only for 1st thread in core */ | 507 | /* print per-core data only for 1st thread in core */ |
@@ -530,79 +509,76 @@ int format_counters(struct thread_data *t, struct core_data *c, | |||
530 | goto done; | 509 | goto done; |
531 | 510 | ||
532 | if (do_nhm_cstates && !do_slm_cstates) | 511 | if (do_nhm_cstates && !do_slm_cstates) |
533 | outp += sprintf(outp, " %6.2f", 100.0 * c->c3/t->tsc); | 512 | outp += sprintf(outp, "%8.2f", 100.0 * c->c3/t->tsc); |
534 | if (do_nhm_cstates) | 513 | if (do_nhm_cstates) |
535 | outp += sprintf(outp, " %6.2f", 100.0 * c->c6/t->tsc); | 514 | outp += sprintf(outp, "%8.2f", 100.0 * c->c6/t->tsc); |
536 | if (do_snb_cstates) | 515 | if (do_snb_cstates) |
537 | outp += sprintf(outp, " %6.2f", 100.0 * c->c7/t->tsc); | 516 | outp += sprintf(outp, "%8.2f", 100.0 * c->c7/t->tsc); |
538 | 517 | ||
539 | if (do_dts) | 518 | if (do_dts) |
540 | outp += sprintf(outp, " %4d", c->core_temp_c); | 519 | outp += sprintf(outp, "%8d", c->core_temp_c); |
541 | 520 | ||
542 | /* print per-package data only for 1st core in package */ | 521 | /* print per-package data only for 1st core in package */ |
543 | if (!(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) | 522 | if (!(t->flags & CPU_IS_FIRST_CORE_IN_PACKAGE)) |
544 | goto done; | 523 | goto done; |
545 | 524 | ||
546 | if (do_ptm) | 525 | if (do_ptm) |
547 | outp += sprintf(outp, " %4d", p->pkg_temp_c); | 526 | outp += sprintf(outp, "%8d", p->pkg_temp_c); |
548 | 527 | ||
549 | if (do_snb_cstates) | 528 | if (do_snb_cstates) |
550 | outp += sprintf(outp, " %6.2f", 100.0 * p->pc2/t->tsc); | 529 | outp += sprintf(outp, "%8.2f", 100.0 * p->pc2/t->tsc); |
551 | if (do_nhm_cstates && !do_slm_cstates) | 530 | if (do_nhm_cstates && !do_slm_cstates) |
552 | outp += sprintf(outp, " %6.2f", 100.0 * p->pc3/t->tsc); | 531 | outp += sprintf(outp, "%8.2f", 100.0 * p->pc3/t->tsc); |
553 | if (do_nhm_cstates && !do_slm_cstates) | 532 | if (do_nhm_cstates && !do_slm_cstates) |
554 | outp += sprintf(outp, " %6.2f", 100.0 * p->pc6/t->tsc); | 533 | outp += sprintf(outp, "%8.2f", 100.0 * p->pc6/t->tsc); |
555 | if (do_snb_cstates) | 534 | if (do_snb_cstates) |
556 | outp += sprintf(outp, " %6.2f", 100.0 * p->pc7/t->tsc); | 535 | outp += sprintf(outp, "%8.2f", 100.0 * p->pc7/t->tsc); |
557 | if (do_c8_c9_c10) { | 536 | if (do_c8_c9_c10) { |
558 | outp += sprintf(outp, " %6.2f", 100.0 * p->pc8/t->tsc); | 537 | outp += sprintf(outp, "%8.2f", 100.0 * p->pc8/t->tsc); |
559 | outp += sprintf(outp, " %6.2f", 100.0 * p->pc9/t->tsc); | 538 | outp += sprintf(outp, "%8.2f", 100.0 * p->pc9/t->tsc); |
560 | outp += sprintf(outp, " %6.2f", 100.0 * p->pc10/t->tsc); | 539 | outp += sprintf(outp, "%8.2f", 100.0 * p->pc10/t->tsc); |
561 | } | 540 | } |
562 | 541 | ||
563 | /* | 542 | /* |
564 | * If measurement interval exceeds minimum RAPL Joule Counter range, | 543 | * If measurement interval exceeds minimum RAPL Joule Counter range, |
565 | * indicate that results are suspect by printing "**" in fraction place. | 544 | * indicate that results are suspect by printing "**" in fraction place. |
566 | */ | 545 | */ |
567 | if (interval_float < rapl_joule_counter_range) { | 546 | if (interval_float < rapl_joule_counter_range) |
568 | fmt5 = " %5.2f"; | 547 | fmt8 = "%8.2f"; |
569 | fmt6 = " %6.2f"; | 548 | else |
570 | } else { | 549 | fmt8 = " %6.0f**"; |
571 | fmt5 = " %3.0f**"; | ||
572 | fmt6 = " %4.0f**"; | ||
573 | } | ||
574 | 550 | ||
575 | if (do_rapl && !rapl_joules) { | 551 | if (do_rapl && !rapl_joules) { |
576 | if (do_rapl & RAPL_PKG) | 552 | if (do_rapl & RAPL_PKG) |
577 | outp += sprintf(outp, fmt6, p->energy_pkg * rapl_energy_units / interval_float); | 553 | outp += sprintf(outp, fmt8, p->energy_pkg * rapl_energy_units / interval_float); |
578 | if (do_rapl & RAPL_CORES) | 554 | if (do_rapl & RAPL_CORES) |
579 | outp += sprintf(outp, fmt6, p->energy_cores * rapl_energy_units / interval_float); | 555 | outp += sprintf(outp, fmt8, p->energy_cores * rapl_energy_units / interval_float); |
580 | if (do_rapl & RAPL_GFX) | 556 | if (do_rapl & RAPL_GFX) |
581 | outp += sprintf(outp, fmt5, p->energy_gfx * rapl_energy_units / interval_float); | 557 | outp += sprintf(outp, fmt8, p->energy_gfx * rapl_energy_units / interval_float); |
582 | if (do_rapl & RAPL_DRAM) | 558 | if (do_rapl & RAPL_DRAM) |
583 | outp += sprintf(outp, fmt5, p->energy_dram * rapl_energy_units / interval_float); | 559 | outp += sprintf(outp, fmt8, p->energy_dram * rapl_energy_units / interval_float); |
584 | if (do_rapl & RAPL_PKG_PERF_STATUS) | 560 | if (do_rapl & RAPL_PKG_PERF_STATUS) |
585 | outp += sprintf(outp, fmt5, 100.0 * p->rapl_pkg_perf_status * rapl_time_units / interval_float); | 561 | outp += sprintf(outp, fmt8, 100.0 * p->rapl_pkg_perf_status * rapl_time_units / interval_float); |
586 | if (do_rapl & RAPL_DRAM_PERF_STATUS) | 562 | if (do_rapl & RAPL_DRAM_PERF_STATUS) |
587 | outp += sprintf(outp, fmt5, 100.0 * p->rapl_dram_perf_status * rapl_time_units / interval_float); | 563 | outp += sprintf(outp, fmt8, 100.0 * p->rapl_dram_perf_status * rapl_time_units / interval_float); |
588 | } else { | 564 | } else { |
589 | if (do_rapl & RAPL_PKG) | 565 | if (do_rapl & RAPL_PKG) |
590 | outp += sprintf(outp, fmt6, | 566 | outp += sprintf(outp, fmt8, |
591 | p->energy_pkg * rapl_energy_units); | 567 | p->energy_pkg * rapl_energy_units); |
592 | if (do_rapl & RAPL_CORES) | 568 | if (do_rapl & RAPL_CORES) |
593 | outp += sprintf(outp, fmt6, | 569 | outp += sprintf(outp, fmt8, |
594 | p->energy_cores * rapl_energy_units); | 570 | p->energy_cores * rapl_energy_units); |
595 | if (do_rapl & RAPL_GFX) | 571 | if (do_rapl & RAPL_GFX) |
596 | outp += sprintf(outp, fmt5, | 572 | outp += sprintf(outp, fmt8, |
597 | p->energy_gfx * rapl_energy_units); | 573 | p->energy_gfx * rapl_energy_units); |
598 | if (do_rapl & RAPL_DRAM) | 574 | if (do_rapl & RAPL_DRAM) |
599 | outp += sprintf(outp, fmt5, | 575 | outp += sprintf(outp, fmt8, |
600 | p->energy_dram * rapl_energy_units); | 576 | p->energy_dram * rapl_energy_units); |
601 | if (do_rapl & RAPL_PKG_PERF_STATUS) | 577 | if (do_rapl & RAPL_PKG_PERF_STATUS) |
602 | outp += sprintf(outp, fmt5, 100.0 * p->rapl_pkg_perf_status * rapl_time_units / interval_float); | 578 | outp += sprintf(outp, fmt8, 100.0 * p->rapl_pkg_perf_status * rapl_time_units / interval_float); |
603 | if (do_rapl & RAPL_DRAM_PERF_STATUS) | 579 | if (do_rapl & RAPL_DRAM_PERF_STATUS) |
604 | outp += sprintf(outp, fmt5, 100.0 * p->rapl_dram_perf_status * rapl_time_units / interval_float); | 580 | outp += sprintf(outp, fmt8, 100.0 * p->rapl_dram_perf_status * rapl_time_units / interval_float); |
605 | outp += sprintf(outp, fmt5, interval_float); | 581 | outp += sprintf(outp, fmt8, interval_float); |
606 | 582 | ||
607 | } | 583 | } |
608 | done: | 584 | done: |
@@ -1516,6 +1492,9 @@ int has_nehalem_turbo_ratio_limit(unsigned int family, unsigned int model) | |||
1516 | case 0x46: /* HSW */ | 1492 | case 0x46: /* HSW */ |
1517 | case 0x37: /* BYT */ | 1493 | case 0x37: /* BYT */ |
1518 | case 0x4D: /* AVN */ | 1494 | case 0x4D: /* AVN */ |
1495 | case 0x3D: /* BDW */ | ||
1496 | case 0x4F: /* BDX */ | ||
1497 | case 0x56: /* BDX-DE */ | ||
1519 | return 1; | 1498 | return 1; |
1520 | case 0x2E: /* Nehalem-EX Xeon - Beckton */ | 1499 | case 0x2E: /* Nehalem-EX Xeon - Beckton */ |
1521 | case 0x2F: /* Westmere-EX Xeon - Eagleton */ | 1500 | case 0x2F: /* Westmere-EX Xeon - Eagleton */ |
@@ -1629,9 +1608,12 @@ void rapl_probe(unsigned int family, unsigned int model) | |||
1629 | case 0x3C: /* HSW */ | 1608 | case 0x3C: /* HSW */ |
1630 | case 0x45: /* HSW */ | 1609 | case 0x45: /* HSW */ |
1631 | case 0x46: /* HSW */ | 1610 | case 0x46: /* HSW */ |
1611 | case 0x3D: /* BDW */ | ||
1632 | do_rapl = RAPL_PKG | RAPL_CORES | RAPL_CORE_POLICY | RAPL_GFX | RAPL_PKG_POWER_INFO; | 1612 | do_rapl = RAPL_PKG | RAPL_CORES | RAPL_CORE_POLICY | RAPL_GFX | RAPL_PKG_POWER_INFO; |
1633 | break; | 1613 | break; |
1634 | case 0x3F: /* HSX */ | 1614 | case 0x3F: /* HSX */ |
1615 | case 0x4F: /* BDX */ | ||
1616 | case 0x56: /* BDX-DE */ | ||
1635 | do_rapl = RAPL_PKG | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_PKG_PERF_STATUS | RAPL_PKG_POWER_INFO; | 1617 | do_rapl = RAPL_PKG | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_PKG_PERF_STATUS | RAPL_PKG_POWER_INFO; |
1636 | break; | 1618 | break; |
1637 | case 0x2D: | 1619 | case 0x2D: |
@@ -1875,6 +1857,9 @@ int is_snb(unsigned int family, unsigned int model) | |||
1875 | case 0x3F: /* HSW */ | 1857 | case 0x3F: /* HSW */ |
1876 | case 0x45: /* HSW */ | 1858 | case 0x45: /* HSW */ |
1877 | case 0x46: /* HSW */ | 1859 | case 0x46: /* HSW */ |
1860 | case 0x3D: /* BDW */ | ||
1861 | case 0x4F: /* BDX */ | ||
1862 | case 0x56: /* BDX-DE */ | ||
1878 | return 1; | 1863 | return 1; |
1879 | } | 1864 | } |
1880 | return 0; | 1865 | return 0; |
@@ -1886,7 +1871,8 @@ int has_c8_c9_c10(unsigned int family, unsigned int model) | |||
1886 | return 0; | 1871 | return 0; |
1887 | 1872 | ||
1888 | switch (model) { | 1873 | switch (model) { |
1889 | case 0x45: | 1874 | case 0x45: /* HSW */ |
1875 | case 0x3D: /* BDW */ | ||
1890 | return 1; | 1876 | return 1; |
1891 | } | 1877 | } |
1892 | return 0; | 1878 | return 0; |
@@ -2455,7 +2441,7 @@ int main(int argc, char **argv) | |||
2455 | cmdline(argc, argv); | 2441 | cmdline(argc, argv); |
2456 | 2442 | ||
2457 | if (verbose) | 2443 | if (verbose) |
2458 | fprintf(stderr, "turbostat v3.6 Dec 2, 2013" | 2444 | fprintf(stderr, "turbostat v3.7 Feb 6, 2014" |
2459 | " - Len Brown <lenb@kernel.org>\n"); | 2445 | " - Len Brown <lenb@kernel.org>\n"); |
2460 | 2446 | ||
2461 | turbostat_init(); | 2447 | turbostat_init(); |