Upload llama-1B/8_GPUS/dp-1_tp-8_pp-1_mbz-8

.gitattributes

@@ -129,3 +129,4 @@ llama-1B/8_GPUS/dp-8_tp-1_pp-1_mbz-2/profiler/ip-26-0-169-86_2283054.17200516115
 llama-1B/8_GPUS/dp-4_tp-1_pp-2_mbz-2/profiler/ip-26-0-169-239_2659139.1720051804197315742.pt.trace.json filter=lfs diff=lfs merge=lfs -text
 llama-1B/8_GPUS/dp-1_tp-8_pp-1_mbz-16/profiler/ip-26-0-171-88_1107214.1720050881924334797.pt.trace.json filter=lfs diff=lfs merge=lfs -text
 llama-1B/8_GPUS/dp-8_tp-1_pp-1_mbz-1/profiler/ip-26-0-171-62_4129089.1720072629336630088.pt.trace.json filter=lfs diff=lfs merge=lfs -text
+llama-1B/8_GPUS/dp-1_tp-8_pp-1_mbz-8/profiler/ip-26-0-171-249_403602.1720133712161162722.pt.trace.json filter=lfs diff=lfs merge=lfs -text

llama-1B/8_GPUS/dp-1_tp-8_pp-1_mbz-8/bench.slurm

@@ -5,7 +5,7 @@
 #SBATCH --partition=hopper-prod
 #SBATCH --nodes=1
 #SBATCH --gres=gpu:8
-#SBATCH --qos=normal
+#SBATCH --qos=low
 #SBATCH --ntasks-per-node=1
 #SBATCH --cpus-per-task=96
 #SBATCH --exclusive

llama-1B/8_GPUS/dp-1_tp-8_pp-1_mbz-8/log_metrics.csv

@@ -0,0 +1,21 @@
+iteration,consumed_tokens,elapsed_time_per_iteration_ms,tokens_per_sec,tokens_per_sec_per_gpu,global_batch_size,lm_loss,lr,model_tflops_per_gpu,hardware_tflops_per_gpu,grad_norm,memory_usage_MiB,peak_allocated_MiB,peak_reserved_MiB
+1,4190000.0000000005,32800.0,128000.0,16000.0,1020.0,11.5,0.0001,145.0,145.0,15.7,2488.82,16891.31,17044.0
+2,8390000.0,25700.0,163000.0,20400.0,1020.0,11.5,9.53e-05,185.0,185.0,16.0,2488.82,16891.31,17044.0
+3,12600000.0,26000.0,161000.0,20100.0,1020.0,12.8,9.05e-05,183.0,183.0,137.0,2488.82,16891.31,17044.0
+4,16800000.0,29700.0,141000.0,17700.0,1020.0,12.2,8.58e-05,160.0,160.0,22.4,2488.76,2488.9,17044.0
+5,21000000.0,29700.0,141000.0,17700.0,1020.0,12.4,8.11e-05,160.0,160.0,43.0,2488.76,16891.31,17044.0
+6,25200000.0,29500.0,142000.0,17700.0,1020.0,11.1,7.63e-05,161.0,161.0,24.7,2488.76,16891.31,17044.0
+7,29400000.0,27600.0,152000.0,19000.0,1020.0,10.2,7.16e-05,172.0,172.0,12.2,2488.76,16891.31,17044.0
+8,33600000.0,26300.0,160000.0,20000.0,1020.0,9.8,6.68e-05,181.0,181.0,7.31,2488.76,16891.31,17044.0
+9,37700000.0,26200.0,160000.0,20000.0,1020.0,9.32,6.21e-05,182.0,182.0,6.66,2488.76,16891.31,17044.0
+10,41900000.0,26100.0,161000.0,20100.0,1020.0,9.22,5.74e-05,182.0,182.0,16.2,2488.76,16891.31,17044.0
+11,46100000.0,26100.0,161000.0,20100.0,1020.0,8.63,5.26e-05,182.0,182.0,7.91,2488.76,16891.31,17044.0
+12,50300000.0,26300.0,159000.0,19900.0,1020.0,8.27,4.79e-05,181.0,181.0,5.43,2488.76,16891.31,17044.0
+13,54500000.0,26000.0,161000.0,20100.0,1020.0,8.1,4.32e-05,183.0,183.0,5.53,2488.76,16891.31,17044.0
+14,58700000.0,25900.0,162000.0,20200.0,1020.0,7.93,3.84e-05,183.0,183.0,5.77,2488.76,16891.31,17044.0
+15,62900000.0,25900.0,162000.0,20200.0,1020.0,7.72,3.37e-05,184.0,184.0,5.16,2488.76,16891.31,17044.0
+16,67099999.99999999,26200.0,160000.0,20000.0,1020.0,7.56,2.89e-05,181.0,181.0,4.91,2488.76,16891.31,17044.0
+17,71300000.0,25900.0,162000.0,20300.0,1020.0,7.45,2.42e-05,184.0,184.0,4.93,2488.76,16891.31,17044.0
+18,75500000.0,25900.0,162000.0,20200.0,1020.0,7.35,1.95e-05,184.0,184.0,4.04,2488.76,16891.31,17044.0
+19,79700000.0,26100.0,161000.0,20100.0,1020.0,7.29,1.47e-05,182.0,182.0,4.12,2488.76,16891.31,17044.0
+20,83900000.0,25900.0,162000.0,20200.0,1020.0,7.23,1e-05,183.0,183.0,3.95,,,

llama-1B/8_GPUS/dp-1_tp-8_pp-1_mbz-8/profiler.csv

@@ -0,0 +1,2 @@
+forward,backward
+1ms 127μs,0ms 340μs

llama-1B/8_GPUS/dp-1_tp-8_pp-1_mbz-8/profiler/ip-26-0-171-249_403602.1720133712161162722.pt.trace.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a68c9ba5df380c192b9b30d76d781ea8da4d28875c2cd06d7e26463e6312de1
+size 18260787421

llama-1B/8_GPUS/dp-1_tp-8_pp-1_mbz-8/status.txt

@@ -1 +1 @@
-timeout
+completed