Kawrakow 86e33fd6f4 Initial Gemma4 support (#1581) ... * Gemma4: WIP * Gemma4: WIP - runs with totally wrong results * Gemma4: WIP - add CPU 512, 512 FA * Gemma4: WIP It gives a meaningful response in llama-cli, but PPL is still much too high. Is this due to tokenizer issues? * Gemma4: this works I had forgotten the softcap on the final output. * Remove log * Gemma4: WIP E4B/E2B * Gemma4: Q4B/E2B appear to work now * gemma4: tokenizer fixes		2026-04-06 10:01:08 +02:00
..
fa	Initial Gemma4 support (#1581)	2026-04-06 10:01:08 +02:00
iqk_common.h	Even better Q4_0 KV cache (CPU) (#1556)	2026-03-31 08:08:34 +02:00
iqk_config.h	Enable AVX-VNNI 256-bit path for Q4_K and Q5_K R4 matmul (#1446)	2026-03-17 10:44:55 +01:00
iqk_cpu_ops.cpp	Log HAVE_FANCY_SIMD via LLAMA_LOG_INFO (#1492)	2026-03-23 08:43:29 +01:00
iqk_cpu_ops.h	Log HAVE_FANCY_SIMD via LLAMA_LOG_INFO (#1492)	2026-03-23 08:43:29 +01:00
iqk_flash_attn.cpp	mtmd: be able to use alternative types for the K*Q multiplication (#1567)	2026-04-02 08:04:05 +02:00
iqk_flash_impl.h	Enable CUDA graphs for MoE models + GPT-OSS support (#689)	2025-08-15 09:18:07 +03:00
iqk_gemm_1bit.cpp	Bonsai support (ARM_NEON) (#1571)	2026-04-03 08:29:15 +02:00
iqk_gemm_1bit.h	Faster iq1_s GEMM via repacking to Q8_0_R8 (#517)	2025-06-11 15:01:34 +03:00
iqk_gemm_floats.cpp	mtmd: be able to use alternative types for the K*Q multiplication (#1567)	2026-04-02 08:04:05 +02:00
iqk_gemm_floats.h	Refactor iqk_mul_mat.cpp (#435)	2025-05-22 10:05:51 +03:00
iqk_gemm_iqk_quants.cpp	AVX512+AVXVNNI GEMM implementation for quants using Q8_K for activations (#710)	2025-08-22 06:27:07 +03:00
iqk_gemm_iqk_quants.h	Much faster CPU prompt processing (part 2) (#533)	2025-06-18 07:29:33 +03:00
iqk_gemm_iquants.cpp	Enable AVX-VNNI 256-bit path for IQ3_XXS and IQ3_S R4 matmul (#1474)	2026-03-20 16:12:07 +01:00
iqk_gemm_iquants.h	IQ2_XXS: much faster CPU prompt processing (#515)	2025-06-11 11:12:30 +03:00
iqk_gemm_kquants.cpp	Enable AVX-VNNI 256-bit path for Q3_K R4 matmul (#1472)	2026-03-23 09:47:29 +01:00
iqk_gemm_kquants.h	Faster CPU prompt processing for Q4_K and Q5_K (#525)	2025-06-13 07:58:15 +03:00
iqk_gemm_ktquants.cpp	Adding IQ1_KT - 1.75 bpw SOTA quants (#616)	2025-07-20 10:05:23 +02:00
iqk_gemm_ktquants.h	Trellis quants: faster CPU prompt processing (#482)	2025-06-01 15:24:05 +03:00
iqk_gemm_legacy_quants.cpp	Optimize mul_mat_q8_1_r8_q8_2 with AVX-512 for faster Q4_K/Q5_K prompt processing (#1578)	2026-04-06 09:07:23 +02:00
iqk_gemm_legacy_quants.h	Much faster CPU prompt processing (part 3) (#534)	2025-06-18 15:30:56 +03:00
iqk_mul_mat.cpp	Initial Gemma4 support (#1581)	2026-04-06 10:01:08 +02:00
iqk_mul_mat.h	Qwen-3.5/Next tweaks (#1447)	2026-03-18 07:32:17 +01:00
iqk_quantize.cpp	Bonsai support (AVX2, generic) (#1570)	2026-04-02 16:54:08 +02:00
iqk_quantize.h	Bonsai support (AVX2, generic) (#1570)	2026-04-02 16:54:08 +02:00
iqk_utils.h	Fix compilation on clang-cl.exe (#1355)	2026-03-04 08:00:28 +01:00