Skip to content

fix(performance): fix API request preview to match actual request #85

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
natoscott merged 3 commits into from
Sep 4, 2026
Merged

fix(performance): fix API request preview to match actual request #85

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
78e8c1b
fix(performance): fix API request preview to match actual request
natoscott Sep 4, 2026
186d32c
refactor(performance): extract shared request body builder for API pr…
natoscott Sep 4, 2026
4f44703
refactor(performance): include GPU-specific fields in request builder
natoscott Sep 4, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 24 additions & 15 deletions app/performance/PerformanceEstimate.tsx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -606,7 +606,7 @@ export default function QuickEstimate() {
memory: {
weight_precision: testWeightPrecision.toLowerCase(),
kv_cache_precision: testKVCachePrecision.toLowerCase(),
gpu_memory_utilization: 0.90
gpu_memory_utilization: currentAicGpu?.gpuMemoryUtilization ?? 0.90
},
gpu: {
gpu_type: gpu,
Expand Down Expand Up @@ -1874,7 +1874,29 @@ export default function QuickEstimate() {
<Button variant="link" isInline onClick={() => setShowApi((s) => !s)}>
{showApi ? 'Hide' : 'Preview'} API request body
</Button>
{showApi && <pre className={styles.apiBody}>{API_PREVIEW}</pre>}
{showApi && (
<pre className={styles.apiBody}>
{JSON.stringify({
model_path: model || '(select model)',
system: gpu || '(select GPU)',
backend: inferenceBackend,
isl: testISL,
osl: testOSL,
batch_size: testConcurrentUsers,
tp_size: testResult?.memory_analysis.tp_size ?? 'auto',
pp_size: testResult?.parallelism_strategy.pp_size ?? testPpSize,
...(testPrefix > 0 && { prefix: testPrefix }),
...(backendVersion && { backend_version: backendVersion }),
...(testWeightPrecision === 'FP8' && { gemm_quant_mode: 'fp8' }),
...(testWeightPrecision === 'INT8' && { gemm_quant_mode: 'int8_wo' }),
...(testWeightPrecision === 'INT4' && { gemm_quant_mode: 'int4_wo' }),
...(testWeightPrecision === 'MXFP4' && { gemm_quant_mode: 'mxfp4' }),
...(testWeightPrecision === 'NVFP4' && { gemm_quant_mode: 'nvfp4' }),
...(testKVCachePrecision === 'FP8' && { kvcache_quant_mode: 'fp8' }),
...(testKVCachePrecision === 'NVFP4' && { kvcache_quant_mode: 'nvfp4' })
Comment thread
coderabbitai[bot] marked this conversation as resolved.
Outdated
}, null, 2)}
</pre>
)}
</div>
</div>
);
Expand All @@ -1895,16 +1917,3 @@ function ConstraintRow({ label, detail, status, term }: { label: string; detail:
</div>
);
}


const API_PREVIEW = `{
"model": { "model_id": "meta-llama/Llama-3.1-8B-Instruct", "max_model_len": "auto" },
"workload": { "isl_tokens": 100, "osl_tokens": 50, "prefix_cache_hit_rate": 0.0,
"requests_per_day": 1000000, "peak_multiplier": 3.0 },
"memory": { "weight_precision": "bf16", "kv_cache_precision": "fp16",
"gpu_memory_utilization": 0.90 },
"hardware": { "gpu_type": "H100_80GB" },
"parallelism": { "tensor_parallel_size": "auto" },
"engine": { "runtime": "vllm", "block_size": 16, "max_num_seqs": 256,
"enable_prefix_caching": true, "enable_chunked_prefill": "auto" }
}`;
Loading
Loading