| /external/python/cpython2/Tools/pybench/ |
| D | pybench.py | 205 warp = 1 variable in Test
219 def __init__(self, warp=None, calibration_runs=None, timer=None): argument
222 if warp is not None:
223 self.rounds = int(self.rounds / warp)
226 self.warp = warp
403 warp = 1 # Warp factor variable in Benchmark
420 def __init__(self, name, verbose=None, timer=None, warp=None, argument
432 if warp is not None:
433 self.warp = warp
489 warp=self.warp,
[all …]
|
| D | README | 70 -w arg set warp factor to arg (10)
123 Running 10 round(s) of the suite at warp factor 10:
246 # 1-2 seconds (at warp 1).
350 100-200ms using warp 10
351 - changed default warp factor to 10 (from 20)
|
| /external/llvm-project/llvm/test/CodeGen/NVPTX/ |
| D | barrier.ll | 3 declare void @llvm.nvvm.bar.warp.sync(i32)
26 ; CHECK: bar.warp.sync [[ID]];
27 call void @llvm.nvvm.bar.warp.sync(i32 %id)
28 ; CHECK: bar.warp.sync 6;
29 call void @llvm.nvvm.bar.warp.sync(i32 6)
|
| /external/skqp/src/compute/hs/gen/ |
| D | target_opencl.c | 67 hsg_transpose_reg_prefix(state->config->warp.lanes_log2), in hsg_transpose_remap()
181 config->warp.lanes_log2, // FIXME - may be different on a SIMD target in hsg_target_opencl()
182 config->warp.lanes_log2, in hsg_target_opencl()
213 for (uint32_t ii=1; ii<=config->warp.lanes_log2; ii++) in hsg_target_opencl()
224 hsg_transpose(config->warp.lanes_log2, in hsg_target_opencl()
305 m->warps * config->warp.lanes, in hsg_target_opencl()
334 m->warps * config->warp.lanes, in hsg_target_opencl()
500 ops->b * config->warp.lanes, in hsg_target_opencl()
508 ops->b * config->warp.lanes, in hsg_target_opencl()
518 ops->b * config->warp.lanes); in hsg_target_opencl()
[all …]
|
| D | target_cuda.c | 67 hsg_transpose_reg_prefix(state->config->warp.lanes_log2), in hsg_transpose_remap()
179 config->warp.lanes_log2, in hsg_target_cuda()
180 config->warp.lanes_log2, in hsg_target_cuda()
212 for (uint32_t ii=1; ii<=config->warp.lanes_log2; ii++) in hsg_target_cuda()
223 hsg_transpose(config->warp.lanes_log2, in hsg_target_cuda()
336 m->warps * config->warp.lanes, in hsg_target_cuda()
373 m->warps * config->warp.lanes, in hsg_target_cuda()
569 ops->b * config->warp.lanes, in hsg_target_cuda()
577 ops->b * config->warp.lanes, in hsg_target_cuda()
587 ops->b * config->warp.lanes); in hsg_target_cuda()
[all …]
|
| D | target_glsl.c | 67 hsg_transpose_reg_prefix(state->config->warp.lanes_log2), in hsg_transpose_remap()
192 config->warp.lanes_log2, // FIXME -- this matters for SIMD in hsg_target_glsl()
193 config->warp.lanes_log2, in hsg_target_glsl()
224 for (uint32_t ii=1; ii<=config->warp.lanes_log2; ii++) in hsg_target_glsl()
235 hsg_transpose(config->warp.lanes_log2, in hsg_target_glsl()
324 m->warps * config->warp.lanes, in hsg_target_glsl()
369 m->warps * config->warp.lanes, in hsg_target_glsl()
588 ops->b * config->warp.lanes, in hsg_target_glsl()
596 ops->b * config->warp.lanes, in hsg_target_glsl()
606 ops->b * config->warp.lanes); in hsg_target_glsl()
[all …]
|
| D | main.c | 149 .warp = {
227 hsg_config.warp.skpw_bs = bs_keys / hsg_merge[0].warps; in hsg_config_init_shared()
240 uint32_t const bs_threads = merge->warps << hsg_config.warp.lanes_log2; in hsg_merge_levels_init_shared()
270 uint32_t const bc_threads = bc_warps_min << hsg_config.warp.lanes_log2; in hsg_merge_levels_init_shared()
728 uint32_t const level = hsg_config.warp.lanes; in hsg_warp_merge()
1014 ops = hsg_warp_half(ops,hsg_config.warp.lanes); in hsg_bs_flip_merge()
1141 ops = hsg_warp_half(ops,hsg_config.warp.lanes); in hsg_bc_clean()
1461 hsg_config.warp.lanes = atoi(optarg); in main()
1462 hsg_config.warp.lanes_log2 = msb_idx_u32(hsg_config.warp.lanes); in main()
|
| D | gen.h | 64 } warp; member
|
| /external/tensorflow/tensorflow/compiler/tf2xla/kernels/ |
| D | resampler_ops.cc | 195 XlaOp BoundSamples(XlaOpKernelContext* ctx, XlaOp warp, in BoundSamples() argument
201 xla::Gt(warp, in BoundSamples()
206 warp, in BoundSamples()
249 XlaOp gather_indices, XlaOp warp, in CalculateGradData() argument
284 reshaped_weights = BoundSamples(ctx, warp, warp_type, warp_shape, in CalculateGradData()
512 XlaOp warp = ctx->Input("warp"); in Compile() local
517 XlaOp top_left = xla::ConvertElementType(warp, xla::S32); in Compile()
526 XlaOp ratio = warp - xla::ConvertElementType(top_left, data_type); in Compile()
555 auto is_ge_zero = xla::Ge(warp, xla::ZerosLike(warp)); in Compile()
558 warp, in Compile()
[all …]
|
| /external/llvm-project/openmp/libomptarget/deviceRTLs/nvptx/docs/ |
| D | ReductionDesign.txt | 82 ReduceData that is already reduced within a warp to a lane in the first
83 warp with minimal shared memory footprint. This is an essential step to
96 On the warp level, we have three versions of the algorithms:
110 algorithm being used here, is set to 0 to signify full warp reduction.
123 An illustration of this algorithm operating on a hypothetical 8-lane full-warp
162 located in a contiguous subset of threads in a warp starting from lane 0.
171 warp woud be:
222 warp). This particular version of shuffle intrinsic we take accepts only
231 where the first half of the (partial) warp is reduced with the second half
232 of the (partial) warp. This is because, the mapping
[all …]
|
| /external/libgav1/libgav1/src/dsp/ |
| D | warp.cc | 433 dsp->warp = Warp_C</*is_compound=*/false, 8, uint8_t>; in Init8bpp()
438 dsp->warp = Warp_C</*is_compound=*/false, 8, uint8_t>; in Init8bpp()
451 dsp->warp = Warp_C</*is_compound=*/false, 10, uint16_t>; in Init10bpp()
456 dsp->warp = Warp_C</*is_compound=*/false, 10, uint16_t>; in Init10bpp()
|
| D | libgav1_dsp.cmake | 71 "${libgav1_source}/dsp/warp.cc"
72 "${libgav1_source}/dsp/warp.h"
|
| D | dsp.h | 854 WarpFunc warp; member
|
| /external/aac/libSBRenc/src/ |
| D | sbrenc_freq_sca.cpp | 428 FIXP_DBL warp = FL2FXCONST_DBL(0.0f); in FDKsbrEnc_UpdateFreqScale() local
450 warp = FL2FXCONST_DBL(0.5f); /* 1.0/(1.0*2.0) */ in FDKsbrEnc_UpdateFreqScale()
452 warp = FL2FXCONST_DBL(1.0f / 2.6f); /* 1.0/(1.3*2.0); */ in FDKsbrEnc_UpdateFreqScale()
460 num_bands1 = numberOfBands(b_p_o, k1, k2, warp); in FDKsbrEnc_UpdateFreqScale()
|
| /external/rust/crates/tokio/ |
| D | README.md | 131 * [`warp`]: A super-easy, composable, web server framework for warp speeds.
146 [`warp`]: https://github.com/seanmonstar/warp
|
| /external/autotest/client/tests/monotonic_time/ |
| D | control | 14 Based on time-warp-test.c by Ingo Molnar.
|
| /external/llvm-project/openmp/libomptarget/deviceRTLs/amdgcn/src/ |
| D | target_impl.hip | 18 // thread's lane number in the warp
27 // thread's lane number in the warp
|
| /external/arm-trusted-firmware/docs/plat/ |
| D | warp7.rst | 88 cp /path/to/linux/arch/boot/dts/imx7s-warp.dtb fiptool_images
104 --hw-config fiptool_images/imx7s-warp.dtb
110 --hw-config fiptool_images/imx7s-warp.dtb \
|
| /external/skqp/src/gpu/effects/ |
| D | GrRRectBlurEffect.fp | 170 // warp the fragment position to the appropriate part of the 9patch blur texture
|
| /external/libgav1/libgav1/src/dsp/arm/ |
| D | warp_neon.cc | 434 dsp->warp = Warp_NEON</*is_compound=*/false>; in Init8bpp()
|
| /external/skqp/src/compute/skc/platforms/cl_12/kernels/ |
| D | fills_expand.cl | 226 // decrement prim_idx by 1 so we can use inclusive warp scan later
|
| D | rasterize.cl | 2275 // iteration of the warp of multi-pixel lines.
2279 // partial warp workloads in the final iteration of each stage. It
2936 // this is a subgroup/warp-centric kernel
3094 // this is a subgroup/warp-centric kernel
3203 // this is a subgroup/warp-centric kernel
3308 // this is a subgroup/warp-centric kernel
|
| /external/libgav1/libgav1/src/dsp/x86/ |
| D | warp_sse4.cc | 505 dsp->warp = Warp_SSE4_1</*is_compound=*/false>; in Init8bpp()
|
| /external/libgav1/ |
| D | Android.bp | 111 "libgav1/src/dsp/warp.cc",
|
| /external/skqp/src/compute/hs/ |
| D | README.md | 290 intra-warp/inter-lane shuffling ― reenable this capability.
|