Re: [Gems-users] error on L2 misses while multiple benchmarks running

[Pau Erola <pauerola@xxxxxxxxxx>]

Yes, it is set to 8. I copy my configuration file below. Thanks for your 
 fast response.
read-configuration /scratch/1/pauerola/checkpoints/8procs/specB_22.conf

instruction-fetch-mode instruction-fetch-trace
istc-disable
dstc-disable
cpu-switch-time 1

echo("Loading Ruby")
load-module ruby

## CONFIG
ruby0.setparam SIMICS_RUBY_MULTIPLIER 1
ruby0.setparam OPAL_RUBY_MULTIPLIER 1

## CHIP
ruby0.setparam g_NUM_PROCESSORS 8
ruby0.setparam g_PROCS_PER_CHIP 8
ruby0.setparam g_NUM_MEMORIES 8
ruby0.setparam_str g_REPLACEMENT_POLICY LRU

## NETWORK
ruby0.setparam NETWORK_LINK_LATENCY 3
ruby0.setparam g_endpoint_bandwidth 1000
ruby0.setparam_str g_NETWORK_TOPOLOGY FILE_SPECIFIED
ruby0.setparam_str g_NETWORK_TOPOLOGY PT_TO_PT
ruby0.setparam NUMBER_OF_VIRTUAL_NETWORKS 6

## L1 - 32k
ruby0.setparam L1_CACHE_ASSOC 4
ruby0.setparam L1_CACHE_NUM_SETS_BITS 7
ruby0.setparam SEQUENCER_TO_CONTROLLER_LATENCY 4
ruby0.setparam L1_RESPONSE_LATENCY 3

## L2 - 512K
ruby0.setparam g_NUM_L2_BANKS 8
ruby0.setparam L2_CACHE_ASSOC 16
ruby0.setparam L2_CACHE_NUM_SETS_BITS 12
ruby0.setparam L2_RESPONSE_LATENCY 12
ruby0.setparam L2_TAG_LATENCY 4
ruby0.setparam L2_REQUEST_LATENCY 12

## MEMORY
ruby0.setparam g_MEMORY_SIZE_BYTES 4294967296
ruby0.setparam MEMORY_RESPONSE_LATENCY_MINUS_2 278
ruby0.setparam g_NUM_MEMORIES 8
ruby0.setparam DIRECTORY_LATENCY 4

ruby0.init
ruby0.load-caches /scratch/1/pauerola/caches/8procs/specB_22.cache

ruby0.clear-stats
c   200000000
print-time cpu0
ruby0.dump-stats /scratch/1/pauerola/stats/8procs/specB_66.unic.stats

Many thanks,
Pau


Mike Marty wrote:
> Is PROCS_PER_CHIP set to 8?
>
> --Mike
>
>
> On Tue, Oct 28, 2008 at 1:44 PM, Pau Erola <pauerola@xxxxxxxxxx 
> <mailto:pauerola@xxxxxxxxxx>> wrote:
>
>     Greetings,
>
>     I'm working with MOESI_CMP_directory protocol on 8core single processor
>     machine with PT_to_PT interconnection network. We are launching 8
>     SPEC2000 benchmarcks on Solaris9 with their corresponding skip-cycles
>     and warmups. With <<con0.input "prstat\n">> we can see them running
>
>     422 root       42M   28M cpu1    59  -20   0:05:13  12% facerec_base.au
>     423 root      211M  197M cpu7    59  -20   0:03:39  12% swim_base.aug16
>     424 root      157M   47M cpu5    59  -20   0:01:13  12% galgel_base.aug
>     425 root     3016K 2520K cpu6    59  -20   0:00:51  11% twolf_base.aug1
>     426 root       28M   27M run     59  -20   0:00:38  11% equake_base.aug
>     427 root     4280K 3736K cpu3    59  -20   0:00:34  10% art_base.aug16a
>     428 root       72M   58M cpu2    59  -20   0:00:34  10% mgrid_base.aug1
>     429 root       96M   96M cpu4    59  -20   0:00:19 7.6% mcf_base.aug16a
>
>
>     After running Ruby for 200M cycles we see that only 4 first cores have
>     L2 misses (512Kb, 16way assoc)
>
>     Total_misses: 1657738
>     total_misses: 1657738 [ 536341 467195 489785 145520 18157 249 286 205 ]
>     user_misses: 1630461 [ 531623 466042 488376 144420 0 0 0 0 ]
>     supervisor_misses: 27277 [ 4718 1153 1409 1100 18157 249 286 205 ]
>
>     instruction_executed: 981861968 [ 26401973 46018520 35723729 79469230
>     194243888 200001684 200000001 200002943 ]
>     simics_cycles_executed: 1600173064 [ 200021633 200021633 200021633
>     200021633 200021633 200021633 200021633 200021633 ]
>     cycles_per_instruction: 1.62973 [ 7.57601 4.34655 5.59913 2.51697
>     1.02974 1.0001 1.00011 1.00009 ]
>     misses_per_thousand_instructions: 1.68836 [ 20.3144 10.1523 13.7104
>     1.83115 0.0934753 0.00124499 0.00143 0.00102498 ]
>
>
>     We have monitored with printf L2 cache misses in Profiler.C and in
>     Sequencer.C and seems that only 0 to 3 cores produce user misses
>
>     2u  0u  3u  1u  2u  0u  1u  3u  0u  2u  1u  3u  2u  1u  0u  3u  2u  1u
>     0u  3u  1u  0u  2u  3u  1u  0u  2u  3u  0u  2u  3u  0u  1u  2u  3u  0u
>     1u  2u  3u  0u  1u  2u  3u  0u  1u  2u  1u  0u  3u  2u  1u  0u  3u  1u
>     2u  0u  3u  1u  0u  2u  3u  0u  2u  3u  0u  2u  3u  1u  0u  2u  3u  0u
>     2u  1u  3u  0u  2u  1u  3u  0u  1u  2u  3u  1u  0u  2u  3u  1u  0u  2u
>     1u  3u  0u  1u  3u  2u  0u  2u  3u  0u  1u  2u  3u  0u  1u  0u  3u  2u
>     1u  0u  2u  3u  1u  0u  2u  1u  3u  0u  1u  3u  2u  0u  1u  3u  2u  0u
>     1u  3u  0u  2u  3u  0u  2u  3u  0u  1u  2u  0u  3u  1u  2u  0u  1u  3u
>
>
>     We have tried to change Solaris scheduler, priorities, launch
>     applications with multiple users,... without successful. It seems that
>     after loading Ruby only 4 cores work.
>
>     Any help will be very grateful.
>
>
>     Many thanks,
>     Pau
>
>
>     _______________________________________________
>     Gems-users mailing list
>     Gems-users@xxxxxxxxxxx <mailto:Gems-users@xxxxxxxxxxx>
>     https://lists.cs.wisc.edu/mailman/listinfo/gems-users
>     Use Google to search the GEMS Users mailing list by adding
>     "site:https://lists.cs.wisc.edu/archive/gems-users/"; to your search.
>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Gems-users mailing list
> Gems-users@xxxxxxxxxxx
> https://lists.cs.wisc.edu/mailman/listinfo/gems-users
> Use Google to search the GEMS Users mailing list by adding "site:https://lists.cs.wisc.edu/archive/gems-users/"; to your search.