mem-ruby,configs: add a configurable GPU L3 to GPU_VIPER #9
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -38,7 +38,7 @@ | |
| import m5 | ||
| from m5.defines import buildEnv | ||
| from m5.objects import * | ||
| from m5.util import addToPath, warn | ||
|
|
||
| from .Ruby import ( | ||
| create_topology, | ||
|
|
@@ -98,7 +98,7 @@ class L2Cache(RubyCache): | |
| def create(self, size, assoc, options): | ||
| self.size = MemorySize(size) | ||
| self.assoc = assoc | ||
| self.replacement_policy = BRRIPRP() | ||
|
|
||
|
|
||
| class CPCntrl(GPU_VIPER_CorePair_Controller, CntrlBase): | ||
|
|
@@ -336,18 +336,23 @@ class L3Cache(RubyCache): | |
| dataArrayBanks = 16 | ||
| tagArrayBanks = 16 | ||
|
|
||
| def create(self, options, ruby_system, system, num_dirs=None): | ||
|
There was a problem hiding this comment. most of the changes here also seem unrelated to adding an L3?
Author
There was a problem hiding this comment. Agreed. The TCC banking changes are not required for the L3. Ill prepare a pr for this |
||
| # num_dirs: number of directory controllers sharing this L3 pool. | ||
| # Defaults to options.num_dirs (CPU-side) if not specified. | ||
| if num_dirs is None: | ||
| num_dirs = options.num_dirs | ||
| self.size = MemorySize(options.l3_size) | ||
| self.size.value /= num_dirs | ||
| self.assoc = options.l3_assoc | ||
| # Each directory controller owns one L3 slice. Configure the internal | ||
| # data and tag bank count of that slice independently of the number of | ||
| # directory controllers. | ||
| self.dataArrayBanks = options.l3_num_banks | ||
| self.tagArrayBanks = options.l3_num_banks | ||
| self.dataAccessLatency = options.l3_data_latency | ||
| self.tagAccessLatency = options.l3_tag_latency | ||
| self.resourceStalls = False | ||
| self.replacement_policy = BRRIPRP() | ||
|
There was a problem hiding this comment. same thing here -- we should use Jarvis' configurable RP code (like @Basemism you did for the ScalarCache) to make this easier to play around with. |
||
|
|
||
|
|
||
| class L3Cntrl(GPU_VIPER_L3Cache_Controller, CntrlBase): | ||
|
|
@@ -382,7 +387,7 @@ def connectWireBuffers( | |
|
|
||
|
|
||
| class DirCntrl(GPU_VIPER_Directory_Controller, CntrlBase): | ||
| def create(self, options, dir_ranges, ruby_system, system, num_dirs=None): | ||
| self.version = self.versionCount() | ||
|
|
||
| self.response_latency = 30 | ||
|
|
@@ -393,7 +398,8 @@ def create(self, options, dir_ranges, ruby_system, system): | |
| ) | ||
|
|
||
| self.L3CacheMemory = L3Cache() | ||
| self.L3CacheMemory.create(options, ruby_system, system, | ||
| num_dirs=num_dirs) | ||
|
|
||
| self.l3_hit_latency = max( | ||
| self.L3CacheMemory.dataAccessLatency, | ||
|
|
@@ -429,6 +435,12 @@ def define_options(parser): | |
| parser.add_argument("--tcp-issue-latency", type=int, default=1) | ||
| parser.add_argument("--l3-data-latency", type=int, default=20) | ||
| parser.add_argument("--l3-tag-latency", type=int, default=15) | ||
| parser.add_argument( | ||
| "--l3-num-banks", | ||
| type=int, | ||
| default=16, | ||
| help="Number of data and tag banks in each directory L3 slice", | ||
| ) | ||
| parser.add_argument("--cpu-to-dir-latency", type=int, default=120) | ||
| parser.add_argument("--gpu-to-dir-latency", type=int, default=120) | ||
| parser.add_argument( | ||
|
|
@@ -438,6 +450,13 @@ def define_options(parser): | |
| "--no-tcc-resource-stalls", action="store_false", default=True | ||
| ) | ||
| parser.add_argument("--use-L3-on-WT", action="store_true", default=False) | ||
| parser.add_argument("--use-gpu-l3", action="store_true", default=False, | ||
| help="Enable L3 (Infinity Cache) fills for GPU " | ||
| "directory controllers") | ||
| parser.add_argument("--l3-exclusive", action="store_true", default=False, | ||
| help="Experimental non-CDNA3 victim-cache policy: " | ||
| "GPU reads consume L3 entries and fills occur " | ||
| "from lower-level evictions") | ||
| parser.add_argument("--num-tbes", type=int, default=256) | ||
| parser.add_argument("--l2-latency", type=int, default=50) # load to use | ||
| parser.add_argument( | ||
|
|
@@ -560,14 +579,14 @@ def construct_dirs(options, system, ruby_system, network): | |
| # For an odd number of CPUs, still create the right number of controllers | ||
| TCC_bits = int(math.log(options.num_tccs, 2)) | ||
|
|
||
| dir_bits = int(math.log(options.num_dirs, 2)) | ||
| block_size_bits = int(math.log(options.cacheline_size, 2)) | ||
| if options.numa_high_bit: | ||
| numa_bit = options.numa_high_bit | ||
| else: | ||
| # if the numa_bit is not specified, set the directory bits as the | ||
| # lowest bits above the block offset bits, and the numa_bit as the | ||
| # highest of those directory bits | ||
| numa_bit = block_size_bits + dir_bits - 1 | ||
|
|
||
| for i in range(options.num_dirs): | ||
|
|
@@ -584,6 +603,7 @@ def construct_dirs(options, system, ruby_system, network): | |
|
|
||
| dir_cntrl = DirCntrl(noTCCdir=True, TCC_select_num_bits=TCC_bits) | ||
| dir_cntrl.create(options, dir_ranges, ruby_system, system) | ||
| dir_cntrl.L3CacheMemory.start_index_bit = block_size_bits + dir_bits | ||
| dir_cntrl.number_of_TBEs = options.num_tbes | ||
| dir_cntrl.useL3OnWT = options.use_L3_on_WT | ||
| dir_cntrl.L2isWB = options.WB_L2 | ||
|
|
@@ -626,6 +646,9 @@ def construct_gpudirs(options, system, ruby_system, network): | |
| dir_cntrl_nodes = [] | ||
| mem_ctrls = [] | ||
|
|
||
| if options.use_gpu_l3: | ||
| warn("GPU L3 write-back is not supported; modeling a write-through L3.") | ||
|
|
||
| xor_low_bit = 0 | ||
|
|
||
| # For an odd number of CPUs, still create the right number of controllers | ||
|
|
@@ -651,9 +674,13 @@ def construct_gpudirs(options, system, ruby_system, network): | |
| TCC_select_num_bits=TCC_bits, | ||
| clk_domain=system.fabric_clk, | ||
| ) | ||
| dir_cntrl.create(options, [addr_range], ruby_system, system, | ||
| num_dirs=options.dgpu_num_dirs) | ||
| dir_cntrl.L3CacheMemory.start_index_bit = block_size_bits + dir_bits | ||
| dir_cntrl.number_of_TBEs = options.num_tbes | ||
| dir_cntrl.GPUonly = True | ||
| dir_cntrl.useL3OnWT = options.use_gpu_l3 | ||
|
There was a problem hiding this comment. @Basemism do you support a WB L3 in this patch? If not, there probably needs to be a warning here.
Author
There was a problem hiding this comment. Not currently, I'll prepare a warning shortly |
||
| dir_cntrl.L3Exclusive = options.l3_exclusive | ||
| dir_cntrl.L2isWB = options.WB_L2 | ||
|
|
||
| # Connect the Directory controller to the ruby network | ||
|
|
||
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
should we update this to use the variant Jarvis integrated into the mainline where the config tells us what RP to use? Then we can just pick BRRIP from there, and it's much more extensible ...