diff options
author | Andreas Hansson <andreas.hansson@arm.com> | 2015-02-03 14:25:52 -0500 |
---|---|---|
committer | Andreas Hansson <andreas.hansson@arm.com> | 2015-02-03 14:25:52 -0500 |
commit | 5ea60a95b3d87fac6723678e07822aed512f966e (patch) | |
tree | a104cbcc0e4f15deb4ad87f6642a16ca1145ab0e /configs | |
parent | 9aad5b45690adae9f1208b6d3cc8dfc670d54543 (diff) | |
download | gem5-5ea60a95b3d87fac6723678e07822aed512f966e.tar.xz |
config: Adjust DRAM channel interleaving defaults
This patch changes the DRAM channel interleaving default behaviour to
be more representative. The default address mapping (RoRaBaCoCh) moves
the channel bits towards the least significant bits, and uses 128 byte
as the default channel interleaving granularity.
These defaults can be overridden if desired, but should serve as a
sensible starting point for most use-cases.
Diffstat (limited to 'configs')
-rw-r--r-- | configs/common/MemConfig.py | 18 |
1 files changed, 11 insertions, 7 deletions
diff --git a/configs/common/MemConfig.py b/configs/common/MemConfig.py index 4692c2d0e..3d1ae4e64 100644 --- a/configs/common/MemConfig.py +++ b/configs/common/MemConfig.py @@ -127,7 +127,7 @@ for alias, target in _mem_aliases_all: # Normal alias _mem_aliases[alias] = target -def create_mem_ctrl(cls, r, i, nbr_mem_ctrls, intlv_bits, cache_line_size): +def create_mem_ctrl(cls, r, i, nbr_mem_ctrls, intlv_bits, intlv_size): """ Helper function for creating a single memoy controller from the given options. This function is invoked multiple times in config_mem function @@ -135,9 +135,7 @@ def create_mem_ctrl(cls, r, i, nbr_mem_ctrls, intlv_bits, cache_line_size): """ import math - # The default behaviour is to interleave on cache line granularity - cache_line_bit = int(math.log(cache_line_size, 2)) - 1 - intlv_low_bit = cache_line_bit + intlv_low_bit = int(math.log(intlv_size, 2)) # Create an instance so we can figure out the address # mapping and row-buffer size @@ -160,13 +158,13 @@ def create_mem_ctrl(cls, r, i, nbr_mem_ctrls, intlv_bits, cache_line_size): rowbuffer_size = ctrl.device_rowbuffer_size.value * \ ctrl.devices_per_rank.value - intlv_low_bit = int(math.log(rowbuffer_size, 2)) - 1 + intlv_low_bit = int(math.log(rowbuffer_size, 2)) # We got all we need to configure the appropriate address # range ctrl.range = m5.objects.AddrRange(r.start, size = r.size(), intlvHighBit = \ - intlv_low_bit + intlv_bits, + intlv_low_bit + intlv_bits - 1, intlvBits = intlv_bits, intlvMatch = i) return ctrl @@ -192,13 +190,19 @@ def config_mem(options, system): cls = get(options.mem_type) mem_ctrls = [] + # The default behaviour is to interleave memory channels on 128 + # byte granularity, or cache line granularity if larger than 128 + # byte. This value is based on the locality seen across a large + # range of workloads. + intlv_size = max(128, system.cache_line_size.value) + # For every range (most systems will only have one), create an # array of controllers and set their parameters to match their # address mapping in the case of a DRAM for r in system.mem_ranges: for i in xrange(nbr_mem_ctrls): mem_ctrl = create_mem_ctrl(cls, r, i, nbr_mem_ctrls, intlv_bits, - system.cache_line_size.value) + intlv_size) # Set the number of ranks based on the command-line # options if it was explicitly set if issubclass(cls, m5.objects.DRAMCtrl) and \ |