Upload llama-1B/8_GPUS/dp-2_tp-4_pp-1_mbz-4

.gitattributes

@@ -130,3 +130,4 @@ llama-1B/8_GPUS/dp-4_tp-1_pp-2_mbz-2/profiler/ip-26-0-169-239_2659139.1720051804
 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-2_tp-4_pp-1_mbz-4/profiler/ip-26-0-163-226_4070653.1720134059871777463.pt.trace.json filter=lfs diff=lfs merge=lfs -text

llama-1B/8_GPUS/dp-2_tp-4_pp-1_mbz-4/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-2_tp-4_pp-1_mbz-4/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
+1,4190000.0000000005,27600.0,152000.0,19000.0,1020.0,11.4,0.0001,172.0,172.0,20.6,3278.7,14544.35,14674.0
+2,8390000.0,22100.0,190000.0,23700.0,1020.0,11.4,9.53e-05,215.0,215.0,20.7,3278.7,14544.35,14674.0
+3,12600000.0,23600.0,178000.0,22200.0,1020.0,11.6,9.05e-05,202.0,202.0,195.0,3278.7,14544.35,14674.0
+4,16800000.0,29100.0,144000.0,18000.0,1020.0,13.6,8.58e-05,163.0,163.0,28.1,3278.64,4601.89,14674.0
+5,21000000.0,29200.0,143000.0,17900.0,1020.0,12.0,8.11e-05,163.0,163.0,48.8,3278.64,14544.35,14674.0
+6,25200000.0,28700.0,146000.0,18300.0,1020.0,10.9,7.63e-05,166.0,166.0,19.7,3278.64,14544.35,14674.0
+7,29400000.0,24300.0,173000.0,21600.0,1020.0,10.4,7.16e-05,196.0,196.0,8.63,3278.64,14544.35,14674.0
+8,33600000.0,23400.0,179000.0,22400.0,1020.0,9.66,6.68e-05,204.0,204.0,6.85,3278.64,14544.35,14674.0
+9,37700000.0,23100.0,181000.0,22700.0,1020.0,11.2,6.21e-05,206.0,206.0,52.6,3278.64,14544.35,14674.0
+10,41900000.0,23100.0,182000.0,22700.0,1020.0,9.07,5.74e-05,206.0,206.0,14.8,3278.64,14544.35,14674.0
+11,46100000.0,23100.0,181000.0,22700.0,1020.0,8.53,5.26e-05,206.0,206.0,6.54,3278.64,14544.35,14674.0
+12,50300000.0,23200.0,181000.0,22600.0,1020.0,8.33,4.79e-05,205.0,205.0,5.78,3278.64,14544.35,14674.0
+13,54500000.0,23100.0,182000.0,22700.0,1020.0,8.1,4.32e-05,206.0,206.0,5.57,3278.64,14544.35,14674.0
+14,58700000.0,23100.0,182000.0,22700.0,1020.0,7.84,3.84e-05,206.0,206.0,5.32,3278.64,14544.35,14674.0
+15,62900000.0,23100.0,181000.0,22700.0,1020.0,7.63,3.37e-05,206.0,206.0,4.84,3278.64,14544.35,14674.0
+16,67099999.99999999,23200.0,181000.0,22600.0,1020.0,7.53,2.89e-05,205.0,205.0,5.04,3278.64,14544.35,14674.0
+17,71300000.0,23100.0,182000.0,22700.0,1020.0,7.45,2.42e-05,206.0,206.0,5.82,3278.64,14544.35,14674.0
+18,75500000.0,23100.0,182000.0,22700.0,1020.0,7.31,1.95e-05,206.0,206.0,5.13,3278.64,14544.35,14674.0
+19,79700000.0,23200.0,181000.0,22600.0,1020.0,7.18,1.47e-05,205.0,205.0,2.98,3278.64,14544.35,14674.0
+20,83900000.0,23100.0,182000.0,22700.0,1020.0,7.14,1e-05,206.0,206.0,3.85,,,

llama-1B/8_GPUS/dp-2_tp-4_pp-1_mbz-4/profiler.csv

@@ -0,0 +1,2 @@
+forward,backward
+1ms 117μs,0ms 335μs

llama-1B/8_GPUS/dp-2_tp-4_pp-1_mbz-4/profiler/ip-26-0-163-226_4070653.1720134059871777463.pt.trace.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13158ea35f4ae76157926376fbfb73055c63aced6859c16d3927195fb7853a02
+size 18365472388

llama-1B/8_GPUS/dp-2_tp-4_pp-1_mbz-4/status.txt

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