| /linux-6.15/drivers/crypto/intel/qat/ |
| H A D | Kconfig | 24 for accelerating crypto and compression workloads.
35 for accelerating crypto and compression workloads.
46 for accelerating crypto and compression workloads.
57 for accelerating crypto and compression workloads.
68 for accelerating crypto and compression workloads.
81 Virtual Function for accelerating crypto and compression workloads.
93 Virtual Function for accelerating crypto and compression workloads.
105 Virtual Function for accelerating crypto and compression workloads.
|
| /linux-6.15/tools/perf/tests/shell/lib/ |
| H A D | perf_metric_validation.py | 13 self.workloads = [wl] # multiple workloads possible
33 .format(self.metric, self.collectedValue, self.workloads,
47 self.workloads = [x for x in workload.split(",") if x]
202 … [TestError([m], self.workloads[self.wlidx], negmetric[m], 0) for m in negmetric.keys()])
332 self.errlist.extend([TestError([name], self.workloads[self.wlidx], val,
344 allres = [{"Workload": self.workloads[i], "Results": self.allresults[i]}
345 for i in range(0, len(self.workloads))]
423 workload = self.workloads[self.wlidx]
542 for i in range(0, len(self.workloads)):
545 self.collect_perf(self.workloads[i])
[all …]
|
| /linux-6.15/tools/perf/Documentation/ |
| H A D | perf-test.txt | 58 Run a built-in workload, to list them use '--list-workloads', current ones include:
68 The datasym and landlock workloads don't accept any.
70 --list-workloads::
71 List the available workloads to use with -w/--workload.
|
| /linux-6.15/drivers/accel/qaic/ |
| H A D | Kconfig | 15 designed to accelerate Deep Learning inference workloads.
18 for users to submit workloads to the devices.
|
| /linux-6.15/Documentation/mm/damon/ |
| H A D | index.rst | 16 of the size of target workloads).
21 their workloads can write personalized applications for better understanding
22 and optimizations of their workloads and systems.
|
| /linux-6.15/drivers/accel/habanalabs/ |
| H A D | Kconfig | 18 designed to accelerate Deep Learning inference and training workloads.
21 the user to submit workloads to the devices.
|
| /linux-6.15/Documentation/gpu/ |
| H A D | drm-compute.rst | 2 Long running workloads and compute
5 Long running workloads (compute) are workloads that will not complete in 10
7 This means that other techniques need to be used to manage those workloads,
|
| H A D | drm-vm-bind-async.rst | 103 exec functions. For long-running workloads, such pipelining of a bind
109 operations for long-running workloads will not allow for pipelining
110 anyway since long-running workloads don't allow for dma-fences as
121 deeply pipelined behind other VM_BIND operations and workloads
|
| /linux-6.15/Documentation/driver-api/ |
| H A D | dma-buf.rst | 292 randomly hangs workloads until the timeout kicks in. Workloads, which from
305 workloads. This also means no implicit fencing for shared buffers in these
327 faults on GPUs are limited to pure compute workloads.
343 - Compute workloads can always be preempted, even when a page fault is pending
346 - DMA fence workloads and workloads which need page fault handling have
349 reservations for DMA fence workloads.
352 hardware resources for DMA fence workloads when they are in-flight. This must
357 all workloads must be flushed from the GPU when switching between jobs
361 made visible anywhere in the system, all compute workloads must be preempted
372 Note that workloads that run on independent hardware like copy engines or other
[all …]
|
| /linux-6.15/Documentation/accel/qaic/ |
| H A D | aic100.rst | 13 inference workloads. They are AI accelerators.
16 (x8). An individual SoC on a card can have up to 16 NSPs for running workloads.
20 performance. AIC100 cards are multi-user capable and able to execute workloads
82 the processors that run the workloads on AIC100. Each NSP is a Qualcomm Hexagon
85 one workload, AIC100 is limited to 16 concurrent workloads. Workload
93 in and out of workloads. AIC100 has one of these. The DMA Bridge has 16
103 This DDR is used to store workloads, data for the workloads, and is used by the
114 for generic compute workloads.
160 ready to process workloads.
245 other workloads.
[all …]
|
| /linux-6.15/Documentation/timers/ |
| H A D | no_hz.rst | 26 workloads, you will normally -not- want this option.
39 right approach, for example, in heavy workloads with lots of tasks
42 hundreds of microseconds). For these types of workloads, scheduling
56 are running light workloads, you should therefore read the following
118 computationally intensive short-iteration workloads: If any CPU is
228 aggressive real-time workloads, which have the option of disabling
230 some workloads will no doubt want to use adaptive ticks to
232 options for these workloads:
252 workloads, which have few such transitions. Careful benchmarking
253 will be required to determine whether or not other workloads
|
| /linux-6.15/drivers/cpuidle/ |
| H A D | Kconfig | 33 Some workloads benefit from using it and it generally should be safe
45 Some virtualized workloads benefit from using it.
|
| /linux-6.15/Documentation/admin-guide/ |
| H A D | workload-tracing.rst | 34 to evaluate safety considerations. We use strace tool to trace workloads.
67 We used strace to trace the perf, stress-ng, paxtest workloads to illustrate
69 be applied to trace other workloads.
101 paxtest workloads to show how to analyze a workload and identify Linux
102 subsystems used by these workloads. Let's start with an overview of these
103 three workloads to get a better understanding of what they do and how to
173 by three workloads we have chose for this analysis.
312 Tracing workloads
315 Now that we understand the workloads, let's start tracing them.
595 information on the resources in use by workloads using strace.
|
| /linux-6.15/drivers/crypto/cavium/nitrox/ |
| H A D | Kconfig | 18 for accelerating crypto workloads.
|
| /linux-6.15/drivers/infiniband/hw/mana/ |
| H A D | Kconfig | 8 for workloads (e.g. DPDK, MPI etc) that uses RDMA verbs to directly
|
| /linux-6.15/Documentation/networking/device_drivers/ethernet/intel/ |
| H A D | idpf.rst | 81 Driver defaults are meant to fit a wide variety of workloads, but if further
89 is tuned for general workloads. The user can customize the interrupt rate
90 control for specific workloads, via ethtool, adjusting the number of
|
| /linux-6.15/Documentation/dev-tools/ |
| H A D | autofdo.rst | 16 for workloads affected by front-end stalls.
23 characteristics similar to the workloads that are intended to be
34 (1) Sample real workloads using a production environment.
|
| /linux-6.15/Documentation/devicetree/bindings/arm/ |
| H A D | blaize.yaml | 24 and a Blaize Graph Streaming Processor for AI and ML workloads,
|
| /linux-6.15/drivers/gpu/drm/i915/gvt/ |
| H A D | scheduler.c | 1331 kmem_cache_destroy(s->workloads); in intel_vgpu_clean_submission()
1423 s->workloads = kmem_cache_create_usercopy("gvt-g_vgpu_workload", in intel_vgpu_setup_submission()
1430 if (!s->workloads) { in intel_vgpu_setup_submission()
1539 kmem_cache_free(s->workloads, workload); in intel_vgpu_destroy_workload()
1548 workload = kmem_cache_zalloc(s->workloads, GFP_KERNEL); in alloc_workload()
1722 kmem_cache_free(s->workloads, workload); in intel_vgpu_create_workload()
1736 kmem_cache_free(s->workloads, workload); in intel_vgpu_create_workload()
1747 kmem_cache_free(s->workloads, workload); in intel_vgpu_create_workload()
|
| /linux-6.15/Documentation/accounting/ |
| H A D | psi.rst | 10 When CPU, memory or IO devices are contended, workloads experience
19 such resource crunches and the time impact it has on complex workloads
23 scarcity aids users in sizing workloads to hardware--or provisioning
|
| /linux-6.15/Documentation/admin-guide/pm/ |
| H A D | intel_uncore_frequency_scaling.rst | 23 Users may have some latency sensitive workloads where they do not want any
24 change to uncore frequency. Also, users may have workloads which require
123 latency sensitive workloads further tuning can be done by SW to
|
| /linux-6.15/tools/perf/tests/ |
| H A D | builtin-test.c | 141 static struct test_workload *workloads[] = { variable
152 for (unsigned i = 0; i < ARRAY_SIZE(workloads) && ({ workload = workloads[i]; 1; }); i++)
|
| /linux-6.15/Documentation/scheduler/ |
| H A D | sched-design-CFS.rst | 104 "server" (i.e., good batching) workloads. It defaults to a setting suitable
105 for desktop workloads. SCHED_BATCH is handled by the CFS scheduler module too.
108 base_slice_ns will have little to no impact on the workloads.
116 than the previous vanilla scheduler: both types of workloads are isolated much
|
| /linux-6.15/security/ |
| H A D | Kconfig.hardening | 178 sees a 1% slowdown, other systems and workloads may vary and you
218 your workloads.
239 workloads have measured as high as 7%.
257 synthetic workloads have measured as high as 8%.
277 workloads. Image size growth depends on architecture, and should
|
| /linux-6.15/Documentation/tools/rtla/ |
| H A D | common_timerlat_options.rst | 48 Set timerlat to run without a workload, and then dispatches user-space workloads
|