Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

pointless micro-optimizations #2127

Open
wants to merge 4 commits into
base: master
Choose a base branch
from
Open

pointless micro-optimizations #2127

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
e52b259
use bitshifts instead of mult for texture lookup
Jaklyy Apr 16, 2024
5b8bd44
very very minor optimization?
Jaklyy Aug 6, 2024
56638c8
Merge branch 'tex-lookup-optimze' into sdfgh
Jaklyy Aug 6, 2024
f6b73de
use bit manip instead of loops for ldm/stm
Jaklyy Aug 8, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
143 changes: 66 additions & 77 deletions src/ARMInterpreter_LoadStore.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -400,11 +400,7 @@ void A_LDM(ARM* cpu)

if (!(cpu->CurInstr & (1<<23)))
{
for (int i = 0; i < 16; i++)
{
if (cpu->CurInstr & (1<<i))
base -= 4;
}
base -= 4 * __builtin_popcount(cpu->CurInstr & 0xFFFF);

if (cpu->CurInstr & (1<<21))
{
Expand All @@ -418,16 +414,17 @@ void A_LDM(ARM* cpu)
if ((cpu->CurInstr & (1<<22)) && !(cpu->CurInstr & (1<<15)))
cpu->UpdateMode(cpu->CPSR, (cpu->CPSR&~0x1F)|0x10, true);

for (int i = 0; i < 15; i++)
u16 reglist = cpu->CurInstr & 0x7FFF;
while (reglist)
{
if (cpu->CurInstr & (1<<i))
{
if (preinc) base += 4;
if (first) cpu->DataRead32 (base, &cpu->R[i]);
else cpu->DataRead32S(base, &cpu->R[i]);
first = false;
if (!preinc) base += 4;
}
int i = __builtin_ctz(reglist);
reglist ^= 1<<i;

if (preinc) base += 4;
if (first) cpu->DataRead32 (base, &cpu->R[i]);
else cpu->DataRead32S(base, &cpu->R[i]);
first = false;
if (!preinc) base += 4;
}

if (cpu->CurInstr & (1<<15))
Expand Down Expand Up @@ -479,11 +476,7 @@ void A_STM(ARM* cpu)

if (!(cpu->CurInstr & (1<<23)))
{
for (u32 i = 0; i < 16; i++)
{
if (cpu->CurInstr & (1<<i))
base -= 4;
}
base -= 4 * __builtin_popcount(cpu->CurInstr & 0xFFFF);

if (cpu->CurInstr & (1<<21))
cpu->R[baseid] = base;
Expand All @@ -503,26 +496,27 @@ void A_STM(ARM* cpu)
cpu->UpdateMode(cpu->CPSR, (cpu->CPSR&~0x1F)|0x10, true);
}

for (u32 i = 0; i < 16; i++)
u16 reglist = cpu->CurInstr & 0xFFFF;
while (reglist)
{
if (cpu->CurInstr & (1<<i))
{
if (preinc) base += 4;
int i = __builtin_ctz(reglist);
reglist ^= 1<<i;

if (i == baseid && !isbanked)
{
if ((cpu->Num == 0) || (!(cpu->CurInstr & ((1<<i)-1))))
first ? cpu->DataWrite32(base, oldbase) : cpu->DataWrite32S(base, oldbase);
else
first ? cpu->DataWrite32(base, base) : cpu->DataWrite32S(base, base); // checkme
}
if (preinc) base += 4;

if (i == baseid && !isbanked)
{
if ((cpu->Num == 0) || (!(cpu->CurInstr & ((1<<i)-1))))
first ? cpu->DataWrite32(base, oldbase) : cpu->DataWrite32S(base, oldbase);
else
first ? cpu->DataWrite32(base, cpu->R[i]) : cpu->DataWrite32S(base, cpu->R[i]);
first ? cpu->DataWrite32(base, base) : cpu->DataWrite32S(base, base); // checkme
}
else
first ? cpu->DataWrite32(base, cpu->R[i]) : cpu->DataWrite32S(base, cpu->R[i]);

first = false;
first = false;

if (!preinc) base += 4;
}
if (!preinc) base += 4;
}

if (cpu->CurInstr & (1<<22))
Expand Down Expand Up @@ -700,31 +694,23 @@ void T_LDR_SPREL(ARM* cpu)

void T_PUSH(ARM* cpu)
{
int nregs = 0;
bool first = true;

for (int i = 0; i < 8; i++)
{
if (cpu->CurInstr & (1<<i))
nregs++;
}

if (cpu->CurInstr & (1<<8))
nregs++;

u32 base = cpu->R[13];
base -= (nregs<<2);
base -= 4 * __builtin_popcount(cpu->CurInstr & 0x1FF);

cpu->R[13] = base;

for (int i = 0; i < 8; i++)
u8 reglist = cpu->CurInstr & 0xFF;
while (reglist)
{
if (cpu->CurInstr & (1<<i))
{
if (first) cpu->DataWrite32 (base, cpu->R[i]);
else cpu->DataWrite32S(base, cpu->R[i]);
first = false;
base += 4;
}
int i = __builtin_ctz(reglist);
reglist ^= 1<<i;

if (first) cpu->DataWrite32 (base, cpu->R[i]);
else cpu->DataWrite32S(base, cpu->R[i]);
first = false;
base += 4;
}

if (cpu->CurInstr & (1<<8))
Expand All @@ -741,15 +727,16 @@ void T_POP(ARM* cpu)
u32 base = cpu->R[13];
bool first = true;

for (int i = 0; i < 8; i++)
u8 reglist = cpu->CurInstr & 0xFF;
while (reglist)
{
if (cpu->CurInstr & (1<<i))
{
if (first) cpu->DataRead32 (base, &cpu->R[i]);
else cpu->DataRead32S(base, &cpu->R[i]);
first = false;
base += 4;
}
int i = __builtin_ctz(reglist);
reglist ^= 1<<i;

if (first) cpu->DataRead32 (base, &cpu->R[i]);
else cpu->DataRead32S(base, &cpu->R[i]);
first = false;
base += 4;
}

if (cpu->CurInstr & (1<<8))
Expand All @@ -771,15 +758,16 @@ void T_STMIA(ARM* cpu)
u32 base = cpu->R[(cpu->CurInstr >> 8) & 0x7];
bool first = true;

for (int i = 0; i < 8; i++)
u8 reglist = cpu->CurInstr & 0xFF;
while (reglist)
{
if (cpu->CurInstr & (1<<i))
{
if (first) cpu->DataWrite32 (base, cpu->R[i]);
else cpu->DataWrite32S(base, cpu->R[i]);
first = false;
base += 4;
}
int i = __builtin_ctz(reglist);
reglist ^= 1<<i;

if (first) cpu->DataWrite32 (base, cpu->R[i]);
else cpu->DataWrite32S(base, cpu->R[i]);
first = false;
base += 4;
}

// TODO: check "Rb included in Rlist" case
Expand All @@ -792,15 +780,16 @@ void T_LDMIA(ARM* cpu)
u32 base = cpu->R[(cpu->CurInstr >> 8) & 0x7];
bool first = true;

for (int i = 0; i < 8; i++)
u8 reglist = cpu->CurInstr & 0xFF;
while (reglist)
{
if (cpu->CurInstr & (1<<i))
{
if (first) cpu->DataRead32 (base, &cpu->R[i]);
else cpu->DataRead32S(base, &cpu->R[i]);
first = false;
base += 4;
}
int i = __builtin_ctz(reglist);
reglist ^= 1<<i;

if (first) cpu->DataRead32 (base, &cpu->R[i]);
else cpu->DataRead32S(base, &cpu->R[i]);
first = false;
base += 4;
}

if (!(cpu->CurInstr & (1<<((cpu->CurInstr >> 8) & 0x7))))
Expand Down
16 changes: 9 additions & 7 deletions src/GPU3D_Soft.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -143,6 +143,8 @@ void SoftRenderer::TextureLookup(const GPU& gpu, u32 texparam, u32 texpal, s16 s
u32 vramaddr = (texparam & 0xFFFF) << 3;

s32 width = 8 << ((texparam >> 20) & 0x7);
// since width is always a multple of 8 (thus a multiple of 2) we can replace all multiplications by width with a bitshift
s32 widthshift = 3 + ((texparam >> 20) & 0x7);
s32 height = 8 << ((texparam >> 23) & 0x7);

s >>= 4;
Expand Down Expand Up @@ -192,7 +194,7 @@ void SoftRenderer::TextureLookup(const GPU& gpu, u32 texparam, u32 texpal, s16 s
{
case 1: // A3I5
{
vramaddr += ((t * width) + s);
vramaddr += ((t << widthshift) + s);
u8 pixel = ReadVRAM_Texture<u8>(vramaddr, gpu);

texpal <<= 4;
Expand All @@ -203,7 +205,7 @@ void SoftRenderer::TextureLookup(const GPU& gpu, u32 texparam, u32 texpal, s16 s

case 2: // 4-color
{
vramaddr += (((t * width) + s) >> 2);
vramaddr += (((t << widthshift) + s) >> 2);
u8 pixel = ReadVRAM_Texture<u8>(vramaddr, gpu);
pixel >>= ((s & 0x3) << 1);
pixel &= 0x3;
Expand All @@ -216,7 +218,7 @@ void SoftRenderer::TextureLookup(const GPU& gpu, u32 texparam, u32 texpal, s16 s

case 3: // 16-color
{
vramaddr += (((t * width) + s) >> 1);
vramaddr += (((t << widthshift) + s) >> 1);
u8 pixel = ReadVRAM_Texture<u8>(vramaddr, gpu);
if (s & 0x1) pixel >>= 4;
else pixel &= 0xF;
Expand All @@ -229,7 +231,7 @@ void SoftRenderer::TextureLookup(const GPU& gpu, u32 texparam, u32 texpal, s16 s

case 4: // 256-color
{
vramaddr += ((t * width) + s);
vramaddr += ((t << widthshift) + s);
u8 pixel = ReadVRAM_Texture<u8>(vramaddr, gpu);

texpal <<= 4;
Expand All @@ -240,7 +242,7 @@ void SoftRenderer::TextureLookup(const GPU& gpu, u32 texparam, u32 texpal, s16 s

case 5: // compressed
{
vramaddr += ((t & 0x3FC) * (width>>2)) + (s & 0x3FC);
vramaddr += ((t & 0x3FC) << (widthshift-2)) + (s & 0x3FC);
vramaddr += (t & 0x3);
vramaddr &= 0x7FFFF; // address used for all calcs wraps around after slot 3

Expand Down Expand Up @@ -352,7 +354,7 @@ void SoftRenderer::TextureLookup(const GPU& gpu, u32 texparam, u32 texpal, s16 s

case 6: // A5I3
{
vramaddr += ((t * width) + s);
vramaddr += ((t << widthshift) + s);
u8 pixel = ReadVRAM_Texture<u8>(vramaddr, gpu);

texpal <<= 4;
Expand All @@ -363,7 +365,7 @@ void SoftRenderer::TextureLookup(const GPU& gpu, u32 texparam, u32 texpal, s16 s

case 7: // direct color
{
vramaddr += (((t * width) + s) << 1);
vramaddr += (((t << widthshift) + s) << 1);
*color = ReadVRAM_Texture<u16>(vramaddr, gpu);
*alpha = (*color & 0x8000) ? 31 : 0;
}
Expand Down
14 changes: 7 additions & 7 deletions src/GPU3D_Soft.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -101,15 +101,15 @@

if ((w0 & 0x1) && !(w1 & 0x1))
{
this->w0n = w0 - 1;
this->w0d = w0 + 1;
this->w1d = w1;
this->w0n = w0 - 1 >> 1;

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 104 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '-'; '-' will be evaluated first [-Wshift-op-parentheses]
this->w0d = w0 + 1 >> 1;

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / x86_64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]

Check warning on line 105 in src/GPU3D_Soft.h

View workflow job for this annotation

GitHub Actions / arm64 

operator '>>' has lower precedence than '+'; '+' will be evaluated first [-Wshift-op-parentheses]
this->w1d = w1 >> 1;
}
else
{
this->w0n = w0 & 0xFFFE;
this->w0d = w0 & 0xFFFE;
this->w1d = w1 & 0xFFFE;
this->w0n = w0 >> 1;
this->w0d = w0 >> 1;
this->w1d = w1 >> 1;
}

this->shift = 9;
Expand Down Expand Up @@ -138,7 +138,7 @@
// this seems to be a proper division on hardware :/
// I haven't been able to find cases that produce imperfect output
if (den == 0) yfactor = 0;
else yfactor = (s32)(num / den);
else yfactor = ((u32)num / den);
}
}

Expand Down
Loading