Conv3d Min Softmax
4 eligible runs
conv
KernelBench level2 problem 24: Conv3d Min Softmax. The computation is the reference PyTorch module's forward pass; the output shape follows the module (mirrored as each implementation's source).
Fastest reported
867.0µs±1.93 · mean of 100 · 1.12× faster than baseline
torch.compile (inductor)PyTorch · MIT · python
Reported evidence · last observed 2026-03-05. Reported by source; not independently reproduced.
Current records
HardwareBest knownImplementationRuns
NVIDIA H100 · env 1867.0 µstorch.compile (inductor)2NVIDIA H100 · env 21.53 mstorch.compile (inductor)2Not measured on B200 for this workload. Challenges →
Source-native comparison · GPU NVIDIA H100 · Workload d = 24 · h = 32 · w = 32 · batch_size = 128 · in_channels = 3 · fp32 · Protocol KernelBench timing scripts · mean · 2 results · last observed 2026-03-05Record history →
Estimated floor 11.3 µs · record 76.94× above itestimate, not evidence ›
DRAM 11.3 µs · bandwidth-bound on H100 SXM
every declared tensor crosses HBM exactly once (3,350 GB/s, H100 SXM datasheet)
no arithmetic formula for this family: bandwidth floor only
headroom-v1: a lower bound from declared tensors and datasheet peaks. A kernel can sit well above it for good reasons.
#
Implementation
Latency
vs #1
Trust
Observed
1torch.compile (inductor)PyTorch867.0µs±1.931.00×Reported · MIT · source2026-03-05stale
1.12× faster than the baseline. Measured exactly what you asked. Reported by source; not independently reproduced.
source mirroredMITno install recipeView source →Run detail →
2969.0µs±0.521.12×Reported · MIT · source2026-03-05stale
Measured exactly what you asked. The source's designated baseline implementation. Reported by source; not independently reproduced.
source mirroredMITno install recipeView source →Run detail →
Implementations
Implementation
Runtime
Best latency
Evidence
Availability
Semantics
Inputs and outputs
input_1fp32 [batch_size, in_channels, d, h, w]
outfloat [out]
Axes and behavior
dvariable
hvariable
wvariable
outvariable
batch_sizevariable
in_channelsvariable
determinismunspecified
constraintsNo mutation or aliasing
Identity
sha2565aa98a99fb32…
Sources: KernelBench baseline timings (2026-03-05) · MITlast observed 2026-03-05How records are decidedJSON