Skip to content

Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦

C

perf: replace ldexpf with arithmetic

TechPizza4db12f632604
diff --git a/c/deepseek_v4.c b/c/deepseek_v4.c
index aab39182..997981ca 100644
--- a/c/deepseek_v4.c
+++ b/c/deepseek_v4.c
@@ -16110,7 +16110,8 @@ static inline __m256 v4_fp8_decode8(__m256i codes) {
_mm256_slli_epi32(_mm256_add_epi32(exp, _mm256_set1_epi32(120)), 23),
_mm256_slli_epi32(man, 20));
__m256 nval = _mm256_castsi256_ps(nbits);
- __m256 sval = _mm256_mul_ps(_mm256_cvtepi32_ps(man), _mm256_set1_ps(ldexpf(1.0f, -9)));
+ float man_factor = 1.0f / (float)(1 << 9);
+ __m256 sval = _mm256_mul_ps(_mm256_cvtepi32_ps(man), _mm256_set1_ps(man_factor));
__m256 is_sub = _mm256_castsi256_ps(_mm256_cmpeq_epi32(exp, _mm256_setzero_si256()));
__m256i sbits = _mm256_or_si256(
_mm256_castps_si256(_mm256_blendv_ps(nval, sval, is_sub)), sgn);