mirror/rpcs3
1
0
Fork 0
mirror of https://github.com/RPCS3/rpcs3.git synced 2025-03-14 10:21:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

Optimize SPU trampolines

Browse Source 
Load values in EAX and reuse it if possible
This commit is contained in:
Nekotekina 2019-02-27 23:09:09 +03:00
parent f143035af1
commit 765d15f23f
2 changed files with 44 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

69
rpcs3/Emu/Cell/SPURecompiler.cpp
View File

@ -341,6 +341,7 @@ void spu_runtime::add(std::pair<const std::vector<u32>, spu_function_t>& where,
workload.emplace_back();
workload.back().size = size0;
workload.back().level = 1;
workload.back().from = 0;
workload.back().rel32 = 0;
workload.back().beg = beg;
workload.back().end = _end;
@ -348,7 +349,7 @@ void spu_runtime::add(std::pair<const std::vector<u32>, spu_function_t>& where,
for (std::size_t i = 0; i < workload.size(); i++)
{
// Get copy of the workload info
spu_runtime::work w = workload[i];
auto w = workload[i];
// Split range in two parts
auto it = w.beg;
@ -357,7 +358,7 @@ void spu_runtime::add(std::pair<const std::vector<u32>, spu_function_t>& where,
u32 size2 = w.size - size1;
std::advance(it2, w.size / 2);
while (true)
while (verify("spu_runtime::work::level overflow" HERE, w.level))
{
it = it2;
size1 = w.size - size2;
@ -431,17 +432,27 @@ void spu_runtime::add(std::pair<const std::vector<u32>, spu_function_t>& where,
}
// Emit 32-bit comparison: cmp [ls+addr], imm32
verify("Asm overflow" HERE), raw + 10 <= wxptr + size0 * 20;
const u32 cmp_lsa = start + (w.level - 1) * 4;
*raw++ = 0x81;
verify("Asm overflow" HERE), raw + 11 <= wxptr + size0 * 20;
if (w.from != w.level)
{
// If necessary (level has advanced), emit load: mov eax, [ls + addr]
#ifdef _WIN32
*raw++ = 0xba;
*raw++ = 0x8b;
*raw++ = 0x82; // ls = rdx
#else
*raw++ = 0xbe;
*raw++ = 0x8b;
*raw++ = 0x86; // ls = rsi
#endif
std::memcpy(raw, &cmp_lsa, 4);
std::memcpy(raw + 4, &x, 4);
raw += 8;
const u32 cmp_lsa = start + (w.level - 1) * 4;
std::memcpy(raw, &cmp_lsa, 4);
raw += 4;
}
// Emit comparison: cmp eax, imm32
*raw++ = 0x3d;
std::memcpy(raw, &x, 4);
raw += 4;
// Low subrange target
if (size1 == 1)
@ -451,10 +462,11 @@ void spu_runtime::add(std::pair<const std::vector<u32>, spu_function_t>& where,
else
{
make_jump(0x82, raw); // jb rel32 (stub)
workload.push_back(w);
workload.back().end = it;
workload.back().size = size1;
workload.back().rel32 = raw;
auto& to = workload.emplace_back(w);
to.end = it;
to.size = size1;
to.rel32 = raw;
to.from = w.level;
}
// Second subrange target
@ -483,10 +495,11 @@ void spu_runtime::add(std::pair<const std::vector<u32>, spu_function_t>& where,
else
{
make_jump(0x87, raw); // ja rel32 (stub)
workload.push_back(w);
workload.back().beg = it2;
workload.back().size = size2;
workload.back().rel32 = raw;
auto& to = workload.emplace_back(w);
to.beg = it2;
to.size = size2;
to.rel32 = raw;
to.from = w.level;
}
const u32 size3 = w.size - size1 - size2;
@ -498,20 +511,22 @@ void spu_runtime::add(std::pair<const std::vector<u32>, spu_function_t>& where,
else
{
make_jump(0xe9, raw); // jmp rel32 (stub)
workload.push_back(w);
workload.back().beg = it;
workload.back().end = it2;
workload.back().size = size3;
workload.back().rel32 = raw;
auto& to = workload.emplace_back(w);
to.beg = it;
to.end = it2;
to.size = size3;
to.rel32 = raw;
to.from = w.level;
}
}
else
{
make_jump(0xe9, raw); // jmp rel32 (stub)
workload.push_back(w);
workload.back().beg = it;
workload.back().size = w.size - size1;
workload.back().rel32 = raw;
auto& to = workload.emplace_back(w);
to.beg = it;
to.size = w.size - size1;
to.rel32 = raw;
to.from = w.level;
}
}
}

3
rpcs3/Emu/Cell/SPURecompiler.h
View File

@ -51,7 +51,8 @@ public:
struct work
{
u32 size;
u32 level;
u16 from;
u16 level;
u8* rel32;
std::map<std::vector<u32>, spu_function_t>::iterator beg;
std::map<std::vector<u32>, spu_function_t>::iterator end;