A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 17, 00:30 UTC
Remove marlin warning (#4918)
Alexander Matveev committed
May 20, 2024 at 10:55 UTC
da5a0b539d6a5fe0c0195513a797814d2c267540
1 file changed
-4
csrc/quantization/gptq_marlin/gptq_marlin.cu
-4
@@ -1519,10 +1519,6 @@ exec_config_t determine_thread_config(int prob_m, int prob_n, int prob_k,
1519
}
1520
}
1521
1522
- printf("WARNING: Marlin kernel is reducing max_m_blocks due to small SM "
1523
- "GPU cache. This may "
1524
- "hurt performance. Consider upgrading your GPU.\n");
1525
-
1522
max_m_blocks--; // Process less M blocks per invocation to reduce cache
1523
// usage
1524
}