piwalker/dolphin
1
0
Fork 0
mirror of https://github.com/dolphin-emu/dolphin.git synced 2025-07-22 22:00:39 -06:00
Code Issues Packages Projects Releases Wiki Activity

Fix CPU Core Count detection and Enable Parallel Shader Compilation

Browse Source 
This does this following things:

- Default to the runtime automatic number of threads for pre-compiling shaders
- Adds a distinct automatic thread count computation for pre-compilation  (which has less other things going on
and should scale better beyond 4 cores)
- Removes the unused logical_core_count field from the CPU detection
- Changes the semantics of num_cores from maximaum addressable number of cores to actually available CPU cores
(which is also how it was actually used)
- Updates the computation of the HTT flag now that AMD no longer lies about it for its Zen processors
- Background shader compilation is *not* enabled by default
This commit is contained in:
DevJPM
2021-01-02 13:47:18 +01:00
committed by unknown
parent 6ca2897d20
commit 61cfd8696e
4 changed files with 20 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
Source/Core/Common/x64CPUDetect.cpp
View File

@ -3,6 +3,7 @@
#include <cstring>
#include <string>
#include <thread>
#include "Common/CPUDetect.h"
#include "Common/CommonTypes.h"
@ -106,7 +107,6 @@ void CPUInfo::Detect()
// Detect family and other misc stuff.
bool ht = false;
HTT = ht;
logical_cpu_count = 1;
if (max_std_fn >= 1)
{
__cpuid(cpu_id, 0x00000001);
@ -120,9 +120,13 @@ void CPUInfo::Detect()
// Detect AMD Zen1, Zen1+ and Zen2
if (family == 23)
bZen1p2 = true;
logical_cpu_count = (cpu_id[1] >> 16) & 0xFF;
ht = (cpu_id[3] >> 28) & 1;
// AMD CPUs before Zen faked this flag and didn't actually
// implement simultaneous multithreading (SMT; Intel calls it HTT)
// but rather some weird middle-ground between 1-2 cores
HTT = ht && (vendor == CPUVendor::Intel || family >= 23);
if ((cpu_id[3] >> 25) & 1)
bSSE = true;
if ((cpu_id[3] >> 26) & 1)
@ -200,35 +204,10 @@ void CPUInfo::Detect()
bLongMode = true;
}
num_cores = (logical_cpu_count == 0) ? 1 : logical_cpu_count;
if (max_ex_fn >= 0x80000008)
{
// Get number of cores. This is a bit complicated. Following AMD manual here.
__cpuid(cpu_id, 0x80000008);
int apic_id_core_id_size = (cpu_id[2] >> 12) & 0xF;
if (apic_id_core_id_size == 0)
{
if (ht)
{
// New mechanism for modern Intel CPUs.
if (vendor == CPUVendor::Intel)
{
__cpuidex(cpu_id, 0x00000004, 0x00000000);
int cores_x_package = ((cpu_id[0] >> 26) & 0x3F) + 1;
HTT = (cores_x_package < logical_cpu_count);
cores_x_package = ((logical_cpu_count % cores_x_package) == 0) ? cores_x_package : 1;
num_cores = (cores_x_package > 1) ? cores_x_package : num_cores;
logical_cpu_count /= cores_x_package;
}
}
}
else
{
// Use AMD's new method.
num_cores = (cpu_id[2] & 0xFF) + 1;
}
}
// this should be much more reliable and easier
// than trying to get the number of cores out of the CPUID data
// ourselves
num_cores = std::max(std::thread::hardware_concurrency(), 1u);
}
// Turn the CPU info into a string we can show