diff options
| author | Andreas Sandberg <Andreas.Sandberg@ARM.com> | 2014-09-03 07:42:22 -0400 |
|---|---|---|
| committer | Andreas Sandberg <Andreas.Sandberg@ARM.com> | 2014-09-03 07:42:22 -0400 |
| commit | 326662b01b0fbb7fe4e38cec7a96222d2891808b (patch) | |
| tree | 35bbca1174a6262d3f69dcf729682e1183f8dede /src/arch/SConscript | |
| parent | e1ac9629398027186ef4c2a66772aeff2b4c6792 (diff) | |
| download | gem5-326662b01b0fbb7fe4e38cec7a96222d2891808b.tar.xz | |
arch, cpu: Factor out the ExecContext into a proper base class
We currently generate and compile one version of the ISA code per CPU
model. This is obviously wasting a lot of resources at compile
time. This changeset factors out the interface into a separate
ExecContext class, which also serves as documentation for the
interface between CPUs and the ISA code. While doing so, this
changeset also fixes up interface inconsistencies between the
different CPU models.
The main argument for using one set of ISA code per CPU model has
always been performance as this avoid indirect branches in the
generated code. However, this argument does not hold water. Booting
Linux on a simulated ARM system running in atomic mode
(opt/10.linux-boot/realview-simple-atomic) is actually 2% faster
(compiled using clang 3.4) after applying this patch. Additionally,
compilation time is decreased by 35%.
Diffstat (limited to 'src/arch/SConscript')
| -rw-r--r-- | src/arch/SConscript | 13 |
1 files changed, 5 insertions, 8 deletions
diff --git a/src/arch/SConscript b/src/arch/SConscript index e9b5c5365..31ea0b78f 100644 --- a/src/arch/SConscript +++ b/src/arch/SConscript @@ -95,13 +95,11 @@ isa_parser = File('isa_parser.py') # The emitter patches up the sources & targets to include the # autogenerated files as targets and isa parser itself as a source. def isa_desc_emitter(target, source, env): - cpu_models = list(env['CPU_MODELS']) - cpu_models.append('CheckerCPU') - # List the isa parser as a source. - source += [ isa_parser ] - # Add in the CPU models. - source += [ Value(m) for m in cpu_models ] + source += [ + isa_parser, + Value("ExecContext"), + ] # Specify different targets depending on if we're running the ISA # parser for its dependency information, or for the generated files. @@ -137,8 +135,7 @@ def isa_desc_action_func(target, source, env): # Skip over the ISA description itself and the parser to the CPU models. models = [ s.get_contents() for s in source[2:] ] - cpu_models = [CpuModel.dict[cpu] for cpu in models] - parser = isa_parser.ISAParser(target[0].dir.abspath, cpu_models) + parser = isa_parser.ISAParser(target[0].dir.abspath) parser.parse_isa_desc(source[0].abspath) isa_desc_action = MakeAction(isa_desc_action_func, Transform("ISA DESC", 1)) |