/* tests/test_bf16_model.c - F32 vs BF16 loader/dispatch parity on a tiny * synthetic model. Hermetic: builds its own model dirs. Runs via * `make test`. */ #include "embed.h" #include "tiny_model.h" #include #include #include #include #include int main(void) { char root[1024], f32_dir[1088], bf16_dir[1088]; snprintf(root, sizeof(root), "%s/embed-bf16-test-XXXXXX", getenv("TMPDIR") ? getenv("TMPDIR") : "/tmp"); if (!mkdtemp(root)) { perror("mkdtemp"); return 2; } snprintf(f32_dir, sizeof(f32_dir), "%s/f32", root); snprintf(bf16_dir, sizeof(bf16_dir), "%s/bf16", root); if (mkdir(f32_dir, 0755) != 0 || mkdir(bf16_dir, 0755) != 0 || tm_write_model(f32_dir, "F32") != 0 || tm_write_model(bf16_dir, "BF16") != 0) { fprintf(stderr, "fixture creation failed\n"); return 2; } int ids[] = {1, 2, 3, 4}; embed_model_t *mf32 = embed_model_load(f32_dir); embed_model_t *mbf16 = embed_model_load(bf16_dir); if (!mf32 || !mbf16) { fprintf(stderr, "model load failed\n"); return 1; } const embed_config_t *cfg = embed_model_config(mf32); int dim = cfg->hidden_size; float *a = (float *)malloc((size_t)dim * sizeof(float)); float *b = (float *)malloc((size_t)dim * sizeof(float)); embed_workspace_t *wf32 = embed_workspace_new(mf32); embed_workspace_t *wbf16 = embed_workspace_new(mbf16); if (!a || !b || !wf32 || !wbf16) { fprintf(stderr, "allocation failed\n"); return 1; } if (embed_model_encode_into(mf32, wf32, ids, 4, a) != 0 || embed_model_encode_into(mbf16, wbf16, ids, 4, b) != 0) { fprintf(stderr, "embedding failed\n"); return 1; } float max_diff = 0.0f; float norm = 0.0f; for (int i = 0; i < dim; i++) { if (!isfinite(a[i]) || !isfinite(b[i])) { fprintf(stderr, "non-finite embedding value\n"); return 1; } float d = fabsf(a[i] - b[i]); if (d > max_diff) max_diff = d; norm += b[i] * b[i]; } norm = sqrtf(norm); /* Embeddings are intentionally unnormalized; require a sane nonzero * magnitude and tight F32-vs-BF16 parity, not unit norm. */ if (norm <= 1e-6f || max_diff > 2e-5f) { fprintf(stderr, "bad parity: norm=%.9g max_diff=%.9g\n", norm, max_diff); return 1; } /* Sequences longer than 16 tokens take the widen-to-F32 weight path * (linear_nobias_weight); parity must hold there too. */ enum { LONG_N = 24 }; int long_ids[LONG_N]; for (int i = 0; i < LONG_N; i++) long_ids[i] = (i % 14) + 1; if (embed_model_encode_into(mf32, wf32, long_ids, LONG_N, a) != 0 || embed_model_encode_into(mbf16, wbf16, long_ids, LONG_N, b) != 0) { fprintf(stderr, "long-sequence embedding failed\n"); return 1; } float long_diff = 0.0f; for (int i = 0; i < dim; i++) { if (!isfinite(a[i]) || !isfinite(b[i])) { fprintf(stderr, "non-finite long-sequence value\n"); return 1; } float d = fabsf(a[i] - b[i]); if (d > long_diff) long_diff = d; } if (long_diff > 2e-5f) { fprintf(stderr, "bad widen-path parity: max_diff=%.9g\n", long_diff); return 1; } printf("ok: bf16 model parity dim=%d max_abs_diff=%.9g norm=%.9g\n", dim, max_diff, norm); embed_workspace_free(wf32); embed_workspace_free(wbf16); embed_model_free(mf32); embed_model_free(mbf16); free(a); free(b); return 0; }