GPUMode MLA Decode
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
GPUMode MLA Decode AMD MI300X Instance 3
1,668.7Decode Latency (µs)
7
GPUMode MLA Decode AMD MI300X Instance 2
1,688.6Runtime (µs)
7
GPUMode MLA Decode AMD MI300X Instance 1
1,653.8Decode Latency (µs)
7