[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Hello Gabriel I have now fixed the issue causing the error. It was related to incorrect generation of the bigCluster and littleCluster. The HelloWorld program is now executing okay when configured to generate the bigCluster and littleCluster. There is one anomaly I wanted to clarify: 1) Running the HelloWorld program, but CHI.py configured for no clusters, I get the following info when I add print statement for cntrl at line 224 in CHI.py. The command used to run being: ./build/ARM/gem5.opt configs/example/se_kirin_custom.py --ruby --topology=Pt2Pt --cpu-type=DerivO3CPU --num-cpus=4 --num-dirs=1 --num-l3caches=1 --num-cpu-bigclust=1 --num-cpu-littleclust=3 --num-clusters=0 --cpu-type-bigclust=DerivO3CPU --cpu-type-littleclust=DerivO3CPU --cmd=tests/test-progs/hello/bin/arm/linux/hello CHI.py -- Cntrl is .cpu0.l1i CHI.py -- Cntrl is .cpu0.l1d CHI.py -- Cntrl is .cpu0.l2 CHI.py -- Cntrl is .cpu1.l1i CHI.py -- Cntrl is .cpu1.l1d CHI.py -- Cntrl is .cpu1.l2 CHI.py -- Cntrl is .cpu2.l1i CHI.py -- Cntrl is .cpu2.l1d CHI.py -- Cntrl is .cpu2.l2 CHI.py -- Cntrl is .cpu3.l1i CHI.py -- Cntrl is .cpu3.l1d CHI.py -- Cntrl is .cpu3.l2 CHI.py -- Cntrl is .ruby.hnf.cntrl CHI.py -- Cntrl is .ruby.snf.cntrl As can be seen the L2 Caches for each CPU is private, as expected. 2) Running the HelloWorld program, but CHI.py configured for [bigCluster, littleCluster], I get the following info when I add print statement for cntrl at line 224 in CHI.py. The command used to run being ./build/ARM/gem5.opt configs/example/se_kirin_custom.py --ruby --topology=Pt2Pt --cpu-type=DerivO3CPU --num-cpus=4 --num-dirs=1 --num-l3caches=1 --num-cpu-bigclust=1 --num-cpu-littleclust=3 --num-clusters=2 --cpu-type-bigclust=DerivO3CPU --cpu-type-littleclust=DerivO3CPU --cmd=tests/test-progs/hello/bin/arm/linux/hello CHI.py -- Cntrl is .cpu0.l1i CHI.py -- Cntrl is .cpu0.l1d CHI.py -- Cntrl is .cpu1.l1i CHI.py -- Cntrl is .cpu1.l1d CHI.py -- Cntrl is .cpu2.l1i CHI.py -- Cntrl is .cpu2.l1d CHI.py -- Cntrl is .cpu0.l2 CHI.py -- Cntrl is .cpu1.l2 CHI.py -- Cntrl is .cpu2.l2 CHI.py -- Cntrl is .cpu3.l1i CHI.py -- Cntrl is .cpu3.l1d CHI.py -- Cntrl is .cpu3.l2 CHI.py -- Cntrl is .ruby.hnf.cntrl CHI.py -- Cntrl is .ruby.snf.cntrl In this case, littleCluster consists of CPU[2:0] and bigCluster consists of CPU[3]. Now for CPU[2:0], the debug info being: CHI.py -- Cntrl is .cpu0.l2 CHI.py -- Cntrl is .cpu1.l2 CHI.py -- Cntrl is .cpu2.l2 Is this showing that the L2Cache is shared between CPU[2:0] or is it showing that each of CPU[0], CPU[1] and CPU[2] has a private L2Cache? Similarly for CPU[3]'s L2 Cache. Thanks in advance Best regards JO -Original Message- From: Javed Osmany Sent: 28 June 2021 12:58 To: gem5 users mailing list Cc: Gabriel Busnot ; Javed Osmany Subject: RE: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hello Gabriel Thank you for the pointer. Best regards J.Osmany -Original Message- From: Gabriel Busnot via gem5-users [mailto:gem5-users@gem5.org] Sent: 28 June 2021 12:49 To: gem5-users@gem5.org<mailto:gem5-users@gem5.org> Cc: Gabriel Busnot mailto:gabriel.bus...@arteris.com>> Subject: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hi Javed, This is an error from the dot file processing tool. For some reason, likely a hard coded string buffer size, it does not support lines longer that 16384 lines. Luckily, you don’t need it and can ignore this error. The second error says that one of the parameters has no set value. In your case, this is data_channel_size in Cache_Controller. See CHI.py line 225 to see what you need to do. Best, Gabriel ___ gem5-users mailing list -- gem5-users@gem5.org<mailto:gem5-users@gem5.org> To unsubscribe send an email to gem5-users-le...@gem5.org<mailto:gem5-users-le...@gem5.org> %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Hello Gabriel Thank you for the pointer. Best regards J.Osmany -Original Message- From: Gabriel Busnot via gem5-users [mailto:gem5-users@gem5.org] Sent: 28 June 2021 12:49 To: gem5-users@gem5.org Cc: Gabriel Busnot Subject: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hi Javed, This is an error from the dot file processing tool. For some reason, likely a hard coded string buffer size, it does not support lines longer that 16384 lines. Luckily, you don’t need it and can ignore this error. The second error says that one of the parameters has no set value. In your case, this is data_channel_size in Cache_Controller. See CHI.py line 225 to see what you need to do. Best, Gabriel ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Hi Javed, This is an error from the dot file processing tool. For some reason, likely a hard coded string buffer size, it does not support lines longer that 16384 lines. Luckily, you don’t need it and can ignore this error. The second error says that one of the parameters has no set value. In your case, this is data_channel_size in Cache_Controller. See CHI.py line 225 to see what you need to do. Best, Gabriel ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Hi Gabriel Outlook blocked the attachments. Resending the two files as WinRar archive. Best regards JO _ From: Javed Osmany Sent: 28 June 2021 11:35 To: gem5 users mailing list Cc: Gabriel Busnot ; Javed Osmany Subject: RE: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hello Gabriel Have tried to follow your flow. Have generated the custom version of CHI.py and se.py (both files attached). I have used HelloWorld program to pipe flush. The command I have used to run being: ./build/ARM/gem5.opt configs/example/se_kirin_custom.py --ruby --topology=Pt2Pt --cpu-type=DerivO3CPU --num-cpus=4 --num-dirs=1 --num-l3caches=1 --num-cpu-bigclust=1 --num-cpu-littleclust=3 --num-clusters=2 --cpu-type-bigclust=DerivO3CPU --cpu-type-littleclust=DerivO3CPU --cmd=tests/test-progs/hello/bin/arm/linux/hello 1. The Ruby memory system seems to have been generated with the two clusters (bigCluster and littleCluster) (checking m5out/config.ini) [system.ruby] type=RubySystem children=bigCluster clk_domain hnf littleCluster network power_state snf access_backing_store=false all_instructions=false block_size_bytes=64 clk_domain=system.ruby.clk_domain eventq_index=0 hot_lines=false memory_size_bits=48 num_of_sequencers=2 number_of_virtual_networks=4 phys_mem=Null power_model= power_state=system.ruby.power_state randomization=false system=system 2. However, I am getting the following run time errors: warn: sys_port_proxy.slave is deprecated. `slave` is now called `in_ports` warn: cntrl.memory is deprecated. The request port for Ruby memory output to the main memory is now called `memory_out_port` Global frequency set at 1 ticks per second warn: system.ruby.network.ext_links0.ext_node adopting orphan SimObject param 'sequencer' warn: system.ruby.network.ext_links1.ext_node adopting orphan SimObject param 'sequencer' ['dot', '-Tsvg', '/tmp/tmp6kqukhxa'] return code: 1 stdout, stderr: b'' b'Error: /tmp/tmp6kqukhxa: syntax error in line 5993 scanning a quoted string (missing endquote? longer than 16384?)\nString starting:"adaptive_routingfalse\\buffer_size4\\clk_domainsystem.rub\n' <== Not sure which file the syntax error is occurring warn: failed to generate dot output from m5out/config.dot warn: DRAM device capacity (8192 Mbytes) does not match the address range assigned (512 Mbytes) fatal: system.cpu0.l1d.data_channel_size without default or user set value Any pointers as to where I should look to resolve the errors? Thanks in advance JO -Original Message- From: Javed Osmany Sent: 24 June 2021 10:09 To: gem5 users mailing list mailto:gem5-users@gem5.org>> Cc: Gabriel Busnot mailto:gabriel.bus...@arteris.com>>; Javed Osmany mailto:javed.osm...@huawei.com>> Subject: RE: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hello Gabriel Thank you for the updated information. Will try to follow for the implementation. Best regards JO -Original Message- From: Gabriel Busnot via gem5-users [mailto:gem5-users@gem5.org] Sent: 24 June 2021 08:33 To: gem5-users@gem5.org<mailto:gem5-users@gem5.org> Cc: Gabriel Busnot mailto:gabriel.bus...@arteris.com>> Subject: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hi Javed, 2- Yes, I meant to define L1ICache_big, L1ICache_little, etc. If you need different cache configurations. However, I didn't get that you need private L2s. But if you call the big and little clusters by the term "cluster", then I believe that each cluster has a single entry point to the coherent interconnect which usually is the master port of a shared cache. In that case, the CHI RN-F is what maps to such concept of CPU cluster. I would recommend starting with private L1s and shared L2 until things work, as this is natively supported by the CHI config files. Then, hack your way through the CHI config files to make the L2 private and add a shared L3 if needed. You will have to modify CHI_config.CHI_RNF. 3- create_system in CHI.py expects system to be populated with CPUs (system.cpu). The only thing that differs between a CPU type and another really is its accuracy : atomic, minor, O3, etc. The performance of the simulated CPU then depends on the specific parameters of that CPU but you can look at that later. Having said that, I would write something like: assert(len(system.cpu) == options.num_cpus_little + options.num_cpus.big) ruby_system.little_cluster = CHI_RNF(system.cpu[:options.num_cpus_little], ) ruby_system.big_cluster = CHI_RNF(system.cpu[-options.num_cpus_big:], ) ruby_system.little_cluster.addPrivL2Cache(...) ruby_system.big_cluster.addPrivL2Cache(...) Best, gabriel ___ gem5-users mailing list -- gem5-users@gem5.org<mailto:gem5-users@gem5.org> To unsubscribe send an email to gem5-users-le...@gem5.org<mailto:gem5-users-le...@gem5.or
[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Hello Gabriel Have tried to follow your flow. Have generated the custom version of CHI.py and se.py (both files attached). I have used HelloWorld program to pipe flush. The command I have used to run being: ./build/ARM/gem5.opt configs/example/se_kirin_custom.py --ruby --topology=Pt2Pt --cpu-type=DerivO3CPU --num-cpus=4 --num-dirs=1 --num-l3caches=1 --num-cpu-bigclust=1 --num-cpu-littleclust=3 --num-clusters=2 --cpu-type-bigclust=DerivO3CPU --cpu-type-littleclust=DerivO3CPU --cmd=tests/test-progs/hello/bin/arm/linux/hello 1. The Ruby memory system seems to have been generated with the two clusters (bigCluster and littleCluster) (checking m5out/config.ini) [system.ruby] type=RubySystem children=bigCluster clk_domain hnf littleCluster network power_state snf access_backing_store=false all_instructions=false block_size_bytes=64 clk_domain=system.ruby.clk_domain eventq_index=0 hot_lines=false memory_size_bits=48 num_of_sequencers=2 number_of_virtual_networks=4 phys_mem=Null power_model= power_state=system.ruby.power_state randomization=false system=system 2. However, I am getting the following run time errors: warn: sys_port_proxy.slave is deprecated. `slave` is now called `in_ports` warn: cntrl.memory is deprecated. The request port for Ruby memory output to the main memory is now called `memory_out_port` Global frequency set at 1 ticks per second warn: system.ruby.network.ext_links0.ext_node adopting orphan SimObject param 'sequencer' warn: system.ruby.network.ext_links1.ext_node adopting orphan SimObject param 'sequencer' ['dot', '-Tsvg', '/tmp/tmp6kqukhxa'] return code: 1 stdout, stderr: b'' b'Error: /tmp/tmp6kqukhxa: syntax error in line 5993 scanning a quoted string (missing endquote? longer than 16384?)\nString starting:"adaptive_routingfalse\\buffer_size4\\clk_domainsystem.rub\n' <== Not sure which file the syntax error is occurring warn: failed to generate dot output from m5out/config.dot warn: DRAM device capacity (8192 Mbytes) does not match the address range assigned (512 Mbytes) fatal: system.cpu0.l1d.data_channel_size without default or user set value Any pointers as to where I should look to resolve the errors? Thanks in advance JO -Original Message- From: Javed Osmany Sent: 24 June 2021 10:09 To: gem5 users mailing list Cc: Gabriel Busnot ; Javed Osmany Subject: RE: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hello Gabriel Thank you for the updated information. Will try to follow for the implementation. Best regards JO -Original Message- From: Gabriel Busnot via gem5-users [mailto:gem5-users@gem5.org] Sent: 24 June 2021 08:33 To: gem5-users@gem5.org<mailto:gem5-users@gem5.org> Cc: Gabriel Busnot mailto:gabriel.bus...@arteris.com>> Subject: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hi Javed, 2- Yes, I meant to define L1ICache_big, L1ICache_little, etc. If you need different cache configurations. However, I didn't get that you need private L2s. But if you call the big and little clusters by the term "cluster", then I believe that each cluster has a single entry point to the coherent interconnect which usually is the master port of a shared cache. In that case, the CHI RN-F is what maps to such concept of CPU cluster. I would recommend starting with private L1s and shared L2 until things work, as this is natively supported by the CHI config files. Then, hack your way through the CHI config files to make the L2 private and add a shared L3 if needed. You will have to modify CHI_config.CHI_RNF. 3- create_system in CHI.py expects system to be populated with CPUs (system.cpu). The only thing that differs between a CPU type and another really is its accuracy : atomic, minor, O3, etc. The performance of the simulated CPU then depends on the specific parameters of that CPU but you can look at that later. Having said that, I would write something like: assert(len(system.cpu) == options.num_cpus_little + options.num_cpus.big) ruby_system.little_cluster = CHI_RNF(system.cpu[:options.num_cpus_little], ) ruby_system.big_cluster = CHI_RNF(system.cpu[-options.num_cpus_big:], ) ruby_system.little_cluster.addPrivL2Cache(...) ruby_system.big_cluster.addPrivL2Cache(...) Best, gabriel ___ gem5-users mailing list -- gem5-users@gem5.org<mailto:gem5-users@gem5.org> To unsubscribe send an email to gem5-users-le...@gem5.org<mailto:gem5-users-le...@gem5.org> %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s CHI.py Description: CHI.py se_kirin_custom.py Description: se_kirin_custom.py ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Hello Gabriel Thank you for the updated information. Will try to follow for the implementation. Best regards JO -Original Message- From: Gabriel Busnot via gem5-users [mailto:gem5-users@gem5.org] Sent: 24 June 2021 08:33 To: gem5-users@gem5.org Cc: Gabriel Busnot Subject: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hi Javed, 2- Yes, I meant to define L1ICache_big, L1ICache_little, etc. If you need different cache configurations. However, I didn't get that you need private L2s. But if you call the big and little clusters by the term "cluster", then I believe that each cluster has a single entry point to the coherent interconnect which usually is the master port of a shared cache. In that case, the CHI RN-F is what maps to such concept of CPU cluster. I would recommend starting with private L1s and shared L2 until things work, as this is natively supported by the CHI config files. Then, hack your way through the CHI config files to make the L2 private and add a shared L3 if needed. You will have to modify CHI_config.CHI_RNF. 3- create_system in CHI.py expects system to be populated with CPUs (system.cpu). The only thing that differs between a CPU type and another really is its accuracy : atomic, minor, O3, etc. The performance of the simulated CPU then depends on the specific parameters of that CPU but you can look at that later. Having said that, I would write something like: assert(len(system.cpu) == options.num_cpus_little + options.num_cpus.big) ruby_system.little_cluster = CHI_RNF(system.cpu[:options.num_cpus_little], ) ruby_system.big_cluster = CHI_RNF(system.cpu[-options.num_cpus_big:], ) ruby_system.little_cluster.addPrivL2Cache(...) ruby_system.big_cluster.addPrivL2Cache(...) Best, gabriel ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Hi Javed, 2- Yes, I meant to define L1ICache_big, L1ICache_little, etc. If you need different cache configurations. However, I didn't get that you need private L2s. But if you call the big and little clusters by the term "cluster", then I believe that each cluster has a single entry point to the coherent interconnect which usually is the master port of a shared cache. In that case, the CHI RN-F is what maps to such concept of CPU cluster. I would recommend starting with private L1s and shared L2 until things work, as this is natively supported by the CHI config files. Then, hack your way through the CHI config files to make the L2 private and add a shared L3 if needed. You will have to modify CHI_config.CHI_RNF. 3- create_system in CHI.py expects system to be populated with CPUs (system.cpu). The only thing that differs between a CPU type and another really is its accuracy : atomic, minor, O3, etc. The performance of the simulated CPU then depends on the specific parameters of that CPU but you can look at that later. Having said that, I would write something like: assert(len(system.cpu) == options.num_cpus_little + options.num_cpus.big) ruby_system.little_cluster = CHI_RNF(system.cpu[:options.num_cpus_little], ) ruby_system.big_cluster = CHI_RNF(system.cpu[-options.num_cpus_big:], ) ruby_system.little_cluster.addPrivL2Cache(...) ruby_system.big_cluster.addPrivL2Cache(...) Best, gabriel ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Hello Gabriel Thank you for the pointer. Made a start on this and just wanted to check if I have understood you correctly.. >> 1- Define two more options in CHI.py to specify the number of big (B) and the number of little (L) cpus from the command line Okay, Done. >> 2- Define the L1ICache, L1DCache and L2DCache for each the big and the little cluster Do you mean for each of the Big and Little CPUs? I originally want the L1$ and L2$ to be private for each CPU. >> 3- Pass the first B cpus as a single list together with the correct caches to the first CHI_RNF. Assign the result to ruby_system.bigCluster. So I tried to following code: Initially assuming that the L1 caches are the same for all the CPU is the system big_cpu = [] for i in range(int(options.num_cpu_bigclust)): big_cpu.append(options.cpu_type_bigclust) print ("Big CPUs are: %s" % (big_cpu)) ## Since number of big cluster cpu is set to 1, it prints out: Big CPUs are: ['DerivO3CPU'] # now define the big cluster # pass the b_cpu list to the CHI_RNF() class ruby_system.bigCluster = [ CHI_RNF([b_cpu], ruby_system, L1ICache, L1DCache, system.cache_line_size.value) for b_cpu in big_cpu ] I am not sure if the above code matches your expectation (high level)??. I am passing the number of big cpus as a list and currently assuming the default L1ICache and L1DCache. Furthermore, the above generates the following error: For both the Big and Little CPU I have used DerivO3CPU Error: could not create sytem for ruby protocol CHI Traceback (most recent call last): File "", line 1, in File "build/ARM/python/m5/main.py", line 455, in main exec(filecode, scope) File "configs/example/se.py", line 254, in Ruby.create_system(options, False, system) File "/home/j00533938/github/gem5-21.0/gem5/configs/ruby/Ruby.py", line 194, in create_system % protocol) File "", line 1, in File "/home/j00533938/github/gem5-21.0/gem5/configs/ruby/CHI.py", line 279, in create_system for b_cpu in big_cpu ] File "/home/j00533938/github/gem5-21.0/gem5/configs/ruby/CHI.py", line 279, in for b_cpu in big_cpu ] <== This is complaining about the above code that I have added, highlighted in purple File "/home/j00533938/github/gem5-21.0/gem5/configs/ruby/CHI_config.py", line 401, in __init__ ruby_system = ruby_system) AttributeError: 'str' object has no attribute 'inst_sequencer' The command line I have used being: ./build/ARM/gem5.opt configs/example/se.py --ruby --topology=Pt2Pt --cpu-type=DerivO3CPU --num-cpus=4 --num-dirs=1 --num-l3caches=1 --num-cpu-bigclust=1 --num-cpu-littleclust=3 --cpu-type-bigclust=DerivO3CPU --cpu-type-littleclust=DerivO3CPU --cmd=tests/test-progs/hello/bin/arm/linux/hello The issue seems to be the following code that I have added: ruby_system.bigCluster = [ CHI_RNF([b_cpu], ruby_system, L1ICache, L1DCache, system.cache_line_size.value) for b_cpu in big_cpu ] Any pointers to help resolve the issue, much appreciated. Best regards JO -Original Message- From: Gabriel Busnot via gem5-users [mailto:gem5-users@gem5.org] Sent: 21 June 2021 17:02 To: gem5-users@gem5.org Cc: Gabriel Busnot Subject: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hi Javed, I don't think that you want to use devices.CpuCluster as it is used to manage classic caches while you want to use Ruby caches. My first approach would be, using se.py as is: 1- Define two more options in CHI.py to specify the number of big (B) and the number of little (L) cpus from the command line 2- Define the L1ICache, L1DCache and L2DCache for each the big and the little cluster 3- Pass the first B cpus as a single list together with the correct caches to the first CHI_RNF. Assign the result to ruby_system.bigCluster. 4- Pass the last L cpus as a single list together with the correct caches to the second CHI_RNF. Assign the result to ruby_system.littleCluster. 5- Add the private L2 cache of the correct type to both cluster. Keep everything else as is. Best, Gabriel ___ gem5-users mailing list -- gem5-users@gem5.org<mailto:gem5-users@gem5.org> To unsubscribe send an email to gem5-users-le...@gem5.org<mailto:gem5-users-le...@gem5.org> %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Many thanks for the pointers, Gabriel. Best Regards J.Osmany -Original Message- From: Gabriel Busnot via gem5-users [mailto:gem5-users@gem5.org] Sent: 21 June 2021 17:02 To: gem5-users@gem5.org Cc: Gabriel Busnot Subject: [gem5-users] Re: [Big, Little] clusters with CHI and SE mode Hi Javed, I don't think that you want to use devices.CpuCluster as it is used to manage classic caches while you want to use Ruby caches. My first approach would be, using se.py as is: 1- Define two more options in CHI.py to specify the number of big (B) and the number of little (L) cpus from the command line 2- Define the L1ICache, L1DCache and L2DCache for each the big and the little cluster 3- Pass the first B cpus as a single list together with the correct caches to the first CHI_RNF. Assign the result to ruby_system.bigCluster. 4- Pass the last L cpus as a single list together with the correct caches to the second CHI_RNF. Assign the result to ruby_system.littleCluster. 5- Add the private L2 cache of the correct type to both cluster. Keep everything else as is. Best, Gabriel ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
[gem5-users] Re: [Big, Little] clusters with CHI and SE mode
Hi Javed, I don't think that you want to use devices.CpuCluster as it is used to manage classic caches while you want to use Ruby caches. My first approach would be, using se.py as is: 1- Define two more options in CHI.py to specify the number of big (B) and the number of little (L) cpus from the command line 2- Define the L1ICache, L1DCache and L2DCache for each the big and the little cluster 3- Pass the first B cpus as a single list together with the correct caches to the first CHI_RNF. Assign the result to ruby_system.bigCluster. 4- Pass the last L cpus as a single list together with the correct caches to the second CHI_RNF. Assign the result to ruby_system.littleCluster. 5- Add the private L2 cache of the correct type to both cluster. Keep everything else as is. Best, Gabriel ___ gem5-users mailing list -- gem5-users@gem5.org To unsubscribe send an email to gem5-users-le...@gem5.org %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s