perf(GFX1100-TG200): T5a vectorizes shared Q8_K quant superblock body - #2822

Open

ghazni101 wants to merge 15 commits into

mudler:mainfrom

ghazni101:row/GFX1100-TG200-T5a

Commits on Sep 6, 2026

feat(KERNEL-QUANT-CIQ-GEMM-ROCM): land the W1 keep-quant providers on kROCM
ghazni101
committed
091a0e6
fix(KERNEL-QUANT-CIQ-GEMM-ROCM): keep grouped Q8_0 on the kROCM provider

mudler
authored and
ghazni101
committed
8dee7ed
Revert "fix(KERNEL-QUANT-CIQ-GEMM-ROCM): keep grouped Q8_0 on the kROCM provider"

mudler
authored and
ghazni101
committed
084e791
fix(KERNEL-QUANT-CIQ-GEMM-ROCM): keep grouped Q8_0 on the kROCM provider

mudler
authored and
ghazni101
committed
1ff81c2
spec(BACKEND-ROCM): define Qwen3.5 gfx1100 numerical gate

VikashLoomba
authored and
ghazni101
committed
b30bc81
spec(BACKEND-ROCM): repair Qwen3.5 numerical plan

VikashLoomba
authored and
ghazni101
committed
88f8a13
spec(BACKEND-ROCM): correct Qwen3.5 capture contracts

VikashLoomba
authored and
ghazni101
committed
9f71bc1
perf(GFX1100-TG200): T4a MMVQ K-quant decode GEMV arm
ghazni101
committed
faef38b
feat(GFX1100-TG200): lever-C fuses Q8_K activation quant into the RmsNorm epilogue
ghazni101
committed
eed5bdf
T27: warp-cooperative QuantizeQ8KK for decode (+2.06%, byte-identical)
ghazni101
committed
c2e6569
feat(GFX1100-TG200): T24 LDS-buffered quant epilogue in RmsNormRowCoopKernel
ghazni101
committed
4163a17
feat(GFX1100-TG200): T21 keep-quant for V-head row-permuted GDN projections
ghazni101
committed
6938253
fix(rocm): gate the coop rmsnorm vector body to bf16 gamma, seed quant amax at 0
ghazni101
committed
e7da1f4
T25: keep ssm_out as Q5_K with runtime input permutation
ghazni101
committed
4f041aa
perf(GFX1100-TG200): T5a vectorizes the shared Q8_K quant superblock body
ghazni101
committed
9f64766

Navigation Menu

perf(GFX1100-TG200): T5a vectorizes shared Q8_K quant superblock body - #2822

perf(GFX1100-TG200): T5a vectorizes shared Q8_K quant superblock body#2822
ghazni101 wants to merge 15 commits into
mudler:mainfrom
ghazni101:row/GFX1100-TG200-T5a

Commits on Sep 6, 2026

feat(KERNEL-QUANT-CIQ-GEMM-ROCM): land the W1 keep-quant providers on kROCM

fix(KERNEL-QUANT-CIQ-GEMM-ROCM): keep grouped Q8_0 on the kROCM provider

Revert "fix(KERNEL-QUANT-CIQ-GEMM-ROCM): keep grouped Q8_0 on the kROCM provider"

fix(KERNEL-QUANT-CIQ-GEMM-ROCM): keep grouped Q8_0 on the kROCM provider

spec(BACKEND-ROCM): define Qwen3.5 gfx1100 numerical gate

spec(BACKEND-ROCM): repair Qwen3.5 numerical plan

spec(BACKEND-ROCM): correct Qwen3.5 capture contracts

perf(GFX1100-TG200): T4a MMVQ K-quant decode GEMV arm

feat(GFX1100-TG200): lever-C fuses Q8_K activation quant into the RmsNorm epilogue

T27: warp-cooperative QuantizeQ8KK for decode (+2.06%, byte-identical)

feat(GFX1100-TG200): T24 LDS-buffered quant epilogue in RmsNormRowCoopKernel

feat(GFX1100-TG200): T21 keep-quant for V-head row-permuted GDN projections

fix(rocm): gate the coop rmsnorm vector body to bf16 gamma, seed quant amax at 0

T25: keep ssm_out as Q5_K with runtime input permutation

perf(GFX1100-TG200): T5a vectorizes the shared Q8_K quant superblock body