Three Optimization Tips for C++

Three Optimization Tips for C++

Andrei Alexandrescu, Ph.D.
Research Scientist, Facebook
andrei.alexandrescu@fb.com

© 2012- Facebook. Do not redistribute. 1 / 33

This Talk

• Basics
• Reduce strength
• Minimize array writes


Things I Shouldn’t Even


Today’s Computing Architectures

• Extremely complex
• Trade reproducible performance for average
speed
• Interrupts, multiprocessing are the norm
• Dynamic frequency control is becoming
common
• Virtually impossible to get identical timings
for experiments


Intuition

• Ignores aspects of a complex reality
• Makes narrow/obsolete/wrong assumptions


Intuition


• “Fewer instructions = faster code”


Intuition


• “Data is faster than computation”


Intuition


• “Computation is faster than data”


Intuition


• “Computation is faster than data”

• The only good intuition: “I should time this.”


Paradox

Measuring gives you a
leg up on experts who
don’t need to measure


Common Pitfalls

• Measuring speed of debug builds
• Different setup for baseline and measured
◦ Sequencing: heap allocator
◦ Warmth of cache, ﬁles, databases, DNS
• Including ancillary work in measurement
◦ malloc, printf common
• Mixtures: measure ta + tb , improve ta ,
conclude tb got improved
• Optimize rare cases, pessimize others


Optimizing Rare Cases


More generalities

• Prefer static linking and PDC
• Prefer 64-bit code, 32-bit data
• Prefer (32-bit) array indexing to pointers
◦ Prefer a[i++] to a[++i]
• Prefer regular memory access patterns
• Minimize ﬂow, avoid data dependencies


Storage Pecking Order

• Use enum for integral constants
• Use static const for other immutables
◦ Beware cache issues
• Use stack for most variables
• Globals: aliasing issues
• thread_local slowest, use local caching
◦ 1 instruction in Windows, Linux
◦ 3-4 in OSX


Reduce Strength


Strength reduction

• Speed hierarchy:
◦ comparisons
◦ (u)int add, subtract, bitops, shift
◦ FP add, sub (separate unit!)
◦ Indexed array access
◦ (u)int32 mul; FP mul
◦ FP division, remainder
◦ (u)int division, remainder


Your Compiler Called

I get it. a >>= 1 is the
same as a /= 2.


Integrals

• Prefer 32-bit ints to all other sizes
◦ 64 bit may make some code slower
◦ 8, 16-bit computations use conversion to
32 bits and back
◦ Use small ints in arrays
• Prefer unsigned to signed
◦ Except when converting to ﬂoating point
• “Most numbers are small”


Floating Point

• Double precision as fast as single precision
• Extended precision just a bit slower
• Do not mix the three
• 1-2 FP addition/subtraction units
• 1-2 FP multiplication/division units
• SSE accelerates throughput for certain
computation kernels
• ints→FPs cheap, FPs→ints expensive


Advice

Design algorithms to
use minimum operation
strength


Strength reduction: Example

• Digit count in base-10 representation
uint32_t digits10(uint64_t v) {
uint32_t result = 0;
do {
++result;
v /= 10;
} while (v);
return result;
}

• Uses integral division extensively
◦ (Actually: multiplication)


Strength reduction: Example

uint32_t result = 1;
for (;;) {
if (v < 10) return result;
if (v < 100) return result + 1;
// Skip ahead by 4 orders of magnitude
v /= 10000U;
result += 4;
}
}

• More comparisons and additions, fewer /=
• (This is not loop unrolling!)

Minimize Array Writes


Minimize Array Writes: Why?

• Disables enregistering
• A write is really a read and a write
• Aliasing makes things difﬁcult
• Maculates the cache

• Generally just difﬁcult to optimize



uint32_t u64ToAsciiClassic(uint64_t value, char* dst) {
// Write backwards.
auto start = dst;
do {
*dst++ = ’0’ + (value % 10);
value /= 10;
} while (value != 0);
const uint32_t result = dst - start;
// Reverse in place.
for (dst--; dst > start; start++, dst--) {
std::iter_swap(dst, start);
}
return result;
}


• Gambit: make one extra pass to compute
length
uint32_t uint64ToAscii(uint64_t v, char *const buffer) {
auto const result = digits10(v);
uint32_t pos = result - 1;
while (v >= 10) {
auto const q = v / 10;
auto const r = static_cast<uint32_t>(v % 10);
buffer[pos--] = ’0’ + r;
v = q;
}
assert(pos == 0);
// Last digit is trivial to handle
*buffer = static_cast<uint32_t>(v) + ’0’;
return result;
}


Improvements

• Fewer array writes
• Regular access patterns
• Fast on small numbers
• Data dependencies reduced


One More Pass

• Reformulate digits10 as search
• Convert two digits at a time


if (v < P01) return 1;
if (v < P12) {
if (v < P08) {
if (v < P06) {
return 5 + (v < P05);
}
return 7 + (v >= P07);
}
if (v < P10) {
return 9 + (v >= P09);
}
return 11 + (v >= P11);
}
return 12 + digits10(v / P12);
}


unsigned u64ToAsciiTable(uint64_t value, char* dst) {
static const char digits[201] =
"0001020304050607080910111213141516171819"
"2021222324252627282930313233343536373839"
"4041424344454647484950515253545556575859"
"6061626364656667686970717273747576777879"
"8081828384858687888990919293949596979899";
uint32_t const length = digits10(value);
uint32_t next = length - 1;
while (value >= 100) {
auto const i = (value % 100) * 2;
value /= 100;
dst[next] = digits[i + 1];
dst[next - 1] = digits[i];
next -= 2;
}


// Handle last 1-2 digits
if (value < 10) {
dst[next] = ’0’ + uint32_t(value);
} else {
auto i = uint32_t(value) * 2;
dst[next] = digits[i + 1];
dst[next - 1] = digits[i];
}
return length;
}


Summary


Summary

• You can’t improve what you can’t measure
◦ Pro tip: You can’t measure what you don’t
measure
• Reduce strength
• Minimize array writes


Three Optimization Tips for C++

Three Optimization Tips for C++

More Related Content

What's hot

Viewers also liked

Similar to Three Optimization Tips for C++

Recently uploaded

Three Optimization Tips for C++