Docs.rs
memra-engine-0.131.0
docs.rs
About docs.rs
Badges
Builds
Metadata
Shorthand URLs
Download
Rustdoc JSON
Build queue
Privacy policy
Rust
Rust website
The Book
Standard Library API Reference
Rust by Example
The Cargo Guide
Clippy Documentation
memra-engine 0.131.0
From-scratch CUDA LLM inference engine for NVIDIA RTX 50-series (sm_120a) and Hopper (sm_90a) - custom kernels, no frameworks
Crate
Source
Builds
Feature flags
..
accprobe_bench.rs
argmax_gate.rs
argmax_margin_probe.rs
b200_matvec_bench.rs
bw_roofline.rs
concat_prime_probe.rs
cpu_native_check.rs
cutlass_gemv_bench.rs
cutlass_smoke.rs
decode_batch_bench.rs
decode_batch_gate.rs
decode_bench.rs
decode_dc_gate.rs
decode_kernel_census.rs
decode_window_profile.rs
depth_profile.rs
dflash2_parity.rs
dflash_parity.rs
dsa_decode_gate.rs
dsa_select_gate.rs
dspark_export_shared.rs
dspark_extract.rs
dspark_generate.rs
dspark_q38_gate.rs
dspark_q38_parity.rs
dspark_sample_gate.rs
dsv4_c4_direct_gate.rs
dsv4_c4_host_gate.rs
dsv4_capacity_gate.rs
dsv4_chunk_sweep_gate.rs
dsv4_coupled_proposal_gate.rs
dsv4_decode_bench.rs
dsv4_decode_probe.rs
dsv4_decode_profile.rs
dsv4_decode_rate_gate.rs
dsv4_drafted_corpus.rs
dsv4_ep_pair_gate.rs
dsv4_gpu_decode_gate.rs
dsv4_gpu_dspark_gate.rs
dsv4_gpu_gate.rs
dsv4_gpu_greedy.rs
dsv4_gpu_sampled_gate.rs
dsv4_gpu_tf_gate.rs
dsv4_graph_probe_gate.rs
dsv4_grouped_prefill_gate.rs
dsv4_hc_grid_gate.rs
dsv4_host_cache_gate.rs
dsv4_http_prompt_pack.rs
dsv4_long_decode_gate.rs
dsv4_matrix_distribution_gate.rs
dsv4_matrix_program_gate.rs
dsv4_native_gemm_gate.rs
dsv4_plain_perf_gate.rs
dsv4_prefill_work_gate.rs
dsv4_sampler_sort_gate.rs
dsv4_sink_score_gate.rs
dsv4_wide_prefill_gate.rs
dtype_gpu_check5.rs
ep_step_fp8_gate.rs
f8f4_check.rs
fa_hd128_check.rs
fa_sanitize.rs
fa_v2_bench.rs
fp8_mmq_bench.rs
fp8_mmq_check.rs
fp8_mmq_stream.rs
frspec_owngen.rs
frspec_rank.rs
fused2_pin.rs
fused_actq_bench.rs
gdn_bench.rs
gemma_gate.rs
gemma_spec_session_gate.rs
gemma_vision_e2e.rs
gemma_vision_oracle.rs
gemv_e4m3_bench.rs
gguf_census.rs
glm5_card3_probe.rs
glm5_decode_graph_gate.rs
glm5_hyper_batch_gate.rs
glm5_hyper_ppn_gate.rs
glm5_prime_v2_gate.rs
glm5_spec_ppn_gate.rs
glm5_tp2_box_probe.rs
glm5_tp_gate.rs
glm5_w8_gate.rs
graph_decode_gate.rs
graph_session_gate.rs
graph_warmup_stress.rs
gu_wall_bench.rs
hc_fused_gate.rs
hy3-layer0-oracle.rs
hy3_ep_kernel_probe.rs
iso_gap_probe.rs
k27div_probe.rs
kernel_check.rs
launch_econ.rs
logit_sample_probe.rs
lt_ndep.rs
m3ref.rs
mla_coalesce_bench.rs
mla_decode_arm_gate.rs
mmvq_bisect.rs
mmvq_pair_m3.rs
modelplan_reference_gate.rs
moe_devq8_check.rs
moe_rows_ceiling_bench.rs
moe_sk_repro.rs
moe_tp2_repro.rs
moe_union_probe.rs
normq_check.rs
nvfp4_bank_oracle.rs
nvfp4_mmq_check.rs
plainbatch_margin_probe.rs
pp2_gate.rs
pp_transport_smoke.rs
ppn_bench.rs
ppn_gate.rs
prime_batch_gate.rs
prime_gate.rs
prime_graph_gate.rs
prime_graph_smoke.rs
q5issue_bench.rs
q8_fuse_gate.rs
quant_kv_f32_check.rs
qwen4exp_gpu_gate.rs
qwen4exp_real_gate.rs
replay_acceptance.rs
router_batch_bench.rs
rp_q4_probe.rs
run_dense.rs
run_eagle.rs
run_gen.rs
run_hybrid.rs
run_lockstep.rs
run_safetensors.rs
run_spec.rs
sample_check.rs
sel_shape_probe.rs
session_bench.rs
session_gate.rs
spec_econ.rs
spec_serve_gate.rs
spec_st_probe.rs
st_vs_gguf.rs
step_prefill_rollback_gate.rs
step_vision_oracle.rs
synth_f8f4.rs
t2probe.rs
target_efficiency.rs
tp2_replicated_row_join_gate.rs
tp_ar_bench.rs
tp_graph_probe.rs
tp_step_attention_gate.rs
tp_step_bf16_gate.rs
tp_step_bf16_row_gate.rs
tp_step_fp8_gate.rs
tp_step_nvfp4_gate.rs
tp_step_resident_device_gate.rs
verify_mscale.rs
verify_probe.rs
verify_shape_bisect.rs
vision_gate.rs
w4a4_gate.rs