Common C++ Performance Mistakes in Games - презентация, доклад, проект скачать

Нажмите для полного просмотра!

Common C++ Performance Mistakes in Games, слайд №1

Common C++ Performance Mistakes in Games, слайд №2

Common C++ Performance Mistakes in Games, слайд №3

Common C++ Performance Mistakes in Games, слайд №4

Common C++ Performance Mistakes in Games, слайд №5

Common C++ Performance Mistakes in Games, слайд №6

Common C++ Performance Mistakes in Games, слайд №7

Common C++ Performance Mistakes in Games, слайд №8

Common C++ Performance Mistakes in Games, слайд №9

Common C++ Performance Mistakes in Games, слайд №10

Common C++ Performance Mistakes in Games, слайд №11

Common C++ Performance Mistakes in Games, слайд №12

Common C++ Performance Mistakes in Games, слайд №13

Common C++ Performance Mistakes in Games, слайд №14

Common C++ Performance Mistakes in Games, слайд №15

Common C++ Performance Mistakes in Games, слайд №16

Common C++ Performance Mistakes in Games, слайд №17

Common C++ Performance Mistakes in Games, слайд №18

Common C++ Performance Mistakes in Games, слайд №19

Common C++ Performance Mistakes in Games, слайд №20

Common C++ Performance Mistakes in Games, слайд №21

Common C++ Performance Mistakes in Games, слайд №22

Common C++ Performance Mistakes in Games, слайд №23

Common C++ Performance Mistakes in Games, слайд №24

Common C++ Performance Mistakes in Games, слайд №25

Common C++ Performance Mistakes in Games, слайд №26

Common C++ Performance Mistakes in Games, слайд №27

Common C++ Performance Mistakes in Games, слайд №28

Common C++ Performance Mistakes in Games, слайд №29

Common C++ Performance Mistakes in Games, слайд №30

Common C++ Performance Mistakes in Games, слайд №31

Common C++ Performance Mistakes in Games, слайд №32

Common C++ Performance Mistakes in Games, слайд №33

Common C++ Performance Mistakes in Games, слайд №34

Common C++ Performance Mistakes in Games, слайд №35

Common C++ Performance Mistakes in Games, слайд №36

Common C++ Performance Mistakes in Games, слайд №37

Common C++ Performance Mistakes in Games, слайд №38

Common C++ Performance Mistakes in Games, слайд №39

Содержание ▲

Вы можете ознакомиться и скачать презентацию на тему Common C++ Performance Mistakes in Games. Доклад-сообщение содержит 39 слайдов. Презентации для любого класса можно скачать бесплатно. Если материал и наш сайт презентаций Mypresentation Вам понравились – поделитесь им с друзьями с помощью социальных кнопок и добавьте в закладки в своем браузере.

Слайды и текст этой презентации

Слайд 1

Описание слайда:

Common C++ Performance Mistakes in Games Pete Isensee Xbox Advanced Technology Group

Слайд 2

Описание слайда:

About the Data ATG reviews code to find bottlenecks and make perf recommendations 50 titles per year 96% use C++ 1 in 3 use “advanced” features like templates or generics

Слайд 3

Описание слайда:

Why This Talk Is Important The majority of Xbox games are CPU bound The CPU bottleneck is often a language or C++ library issue These issues are not usually specific to the platform

Слайд 4

Описание слайда:

Format Definition of the problem Examples Recommendation For reference A frame is 17 or 33 ms (60fps / 30fps) Bottlenecks given in ms per frame

Слайд 5

Описание слайда:

Issue: STL Game using std::list Adding ~20,000 objects every frame Rebuilding the list every frame Time spent: 6.5 ms/frame! ~156K overhead (2 pointers per node) Objects spread all over the heap

Слайд 6

Описание слайда:

std::set and map Many games use set/map as sorted lists Inserts are slow (log(N)) Memory overhead: 3 ptrs + color Worst case in game: 3.8 ms/frame

Слайд 7

Описание слайда:

std::vector Hundreds of push_back()s per frame VS7.1 expands vector by 50% Question: How many reallocations for 100 push_back()s? Answer: 13! (1,2,3,4,5,7,10,14,20,29,43,64,95)

Слайд 8

Описание слайда:

Clearly, the STL is Evil

Слайд 9

Описание слайда:

Use the Right Tool for the Job The STL is powerful, but it’s not free Filling any container is expensive Be aware of container overhead Be aware of heap fragmentation and cache coherency Prefer vector, vector::reserve()

Слайд 10

Описание слайда:

The STL is Evil, Sometimes The STL doesn’t solve every problem The STL solves some problems poorly Sometimes good old C-arrays are the perfect container Mike Abrash puts it well: “The best optimizer is between your ears”

Слайд 11

Описание слайда:

Issue: NIH Syndrome Example: Custom binary tree Sorted list of transparent objects Badly unbalanced 1 ms/frame to add only 400 items Example: Custom dynamic array class Poorer performance than std::vector Fewer features

Слайд 12

$Optimizations that Aren’t void appMemcpy( void* d, const void* s, size_t b ) { // lots of assembly code here ... } appMemcpy( pDest, pSrc, 100 ); //...$

Описание слайда:

Optimizations that Aren’t void appMemcpy( void* d, const void* s, size_t b ) { // lots of assembly code here ... } appMemcpy( pDest, pSrc, 100 ); // bottleneck appMemcpy was slower than memcpy for anything under 64K

Слайд 13

Описание слайда:

Invent Only What You Need std::set/map more efficient than the custom tree by 10X Tested and proven Still high overhead An even better solution Unsorted vector or array Sort once 20X improvement

Слайд 14

Описание слайда:

Profile Run your profiler Rinse. Repeat. Prove the improvement. Don’t rewrite the C runtime or STL just because you can. There are more interesting places to spend your time.

Слайд 15

Описание слайда:

Issue: Tool Knowledge If you’re a programmer, you use C/C++ every day C++ is complex CRT and STL libraries are complex The complexities matter Sometimes they really matter

Слайд 16

Описание слайда:

vector::clear Game reused global vector in frame loop clear() called every frame to empty the vector C++ Standard clear() erases all elements (size() goes to 0) No mention of what happens to vector capacity On VS7.1/Dinkumware, frees the memory Every frame reallocated memory

Слайд 17

Описание слайда:

Zero-Initialization struct Array { int x[1000]; }; struct Container { Array arr; Container() : arr() { } }; Container x; // bottleneck Costing 3.5 ms/frame Removing : arr() speeds this by 20X

Слайд 18

Описание слайда:

Know Thine Holy Standard Use resize(0) to reduce container size without affecting capacity T() means zero-initialize PODs. Don’t use T() unless you mean it. Get a copy of the C++ Standard. Really. search on 14882 Only $18 for the PDF

Слайд 19

Описание слайда:

Issue: C Runtime void BuildScore( char* s, int n ) { if( n > 0 ) sprintf( s, “%d”, n ); else sprintf( s, “” ); } n was often zero sprintf was a hotspot

Слайд 20

Описание слайда:

qsort Sorting is important in games qsort is not an ideal sorting function No type safety Comparison function call overhead No opportunity for compiler inlining There are faster options

Слайд 21

Описание слайда:

Clearly, the CRT is Evil

Слайд 22

$Understand Your Options itoa() can replace sprintf( s, “%d”, n ) *s = ‘\0’ can replace sprintf( s, “” ) std::sort can replace qsort Type safe...$

Описание слайда:

Understand Your Options itoa() can replace sprintf( s, “%d”, n ) *s = ‘\0’ can replace sprintf( s, “” ) std::sort can replace qsort Type safe Comparison can be inlined Other sorting options can be even faster: partial_sort, partition

Слайд 23

Описание слайда:

Issue: Function Calls 50,000-100,000 calls/frame is normal At 60Hz, Xbox has 12.2M cycles/frame Function call/return averages 20 cycles A game calling 61,000 functions/frame spends 10% CPU (1.7 ms/frame) in function call overhead

Слайд 24

Описание слайда:

Extreme Function-ality 120,000 functions/frame 140,000 functions/frame 130,000 calls to a single function/frame (ColumnVec::operator[]) And the winner: 340,000 calls per frame! 9 ms/frame of call overhead

Слайд 25

Описание слайда:

Beware Elegance Elegance → levels of indirection → more functions → perf impact Use algorithmic solutions first One pass through the world Better object rejection Do AI/physics/networking less often than once/frame

Слайд 26

Описание слайда:

Inline Judiciously Remember: inline is a suggestion Try “inline any suitable” compiler option 15 to 20 fps 68,000 calls down to 47,000 Try __forceinline or similar keyword Adding to 5 funcs shaved 1.5 ms/frame Don’t over-inline

Слайд 27

Описание слайда:

Issue: for loops // Example 1: Copy indices to push buffer for( DWORD i = 0; i < dwIndexCnt; ++i ) *pPushBuffer++ = arrIndices[ i ]; // Example 2: Initialize vector array for( DWORD i = 0; i < dwMax; ++i ) mVectorArr[i] = XGVECTOR4(0,0,0,0); // Example 3: Process items in world for( itr i = c.begin(); i < c.end(); ++i ) Process( *i );

Слайд 28

Описание слайда:

Watch Out For For Never copy/clear a POD with a for loop std::algorithms are optimized; use them memcpy( pPushBuffer, arrIndices, dwIndexCnt * sizeof(DWORD) ); memset( mVectorArr, 0, dwMax * sizeof(XGVECTOR4) ); for_each( c.begin(), c.end(), Process );

Слайд 29

Описание слайда:

Issue: Exception Handling Most games never throw Most games never catch Yet, most games enable EH EH adds code to do stack unwinding A little bit of overhead to a lot of code 10% size increase is common 2 ms/frame in worst case

Слайд 30

Описание слайда:

Disable Exception Handling Don’t throw or catch exceptions Turn off the C++ EH compiler option For Dinkumware STL Define “_HAS_EXCEPTIONS=0” Write empty _Throw and _Raise_handler; see stdthrow.cpp and raisehan.cpp in crt folder Add #pragma warning(disable: 4530)

Слайд 31

Описание слайда:

Issue: Strings Programmers love strings Love hurts ~7000 calls to stricmp in frame loop 1.5 ms/frame Binary search of a string table 2 ms/frame

Слайд 32

Описание слайда:

Avoid strings String comparisons don’t belong in the frame loop Put strings in an table and compare indices At least optimize the comparison Compare pointers only Prefer strcmp to stricmp

Слайд 33

Описание слайда:

Issue: Memory Allocation Memory overhead Xbox granularity/overhead is 16/16 bytes Overhead alone is often 1+ MB Too many allocations Games commonly do thousands of allocations per frame Cost: 1-5 ms/frame

Слайд 34

Описание слайда:

Hidden Allocations push_back(), insert() and friends typically allocate memory String constructors allocate Init-style calls often allocate Temporary objects, particularly string constants that convert to string objects

Слайд 35

Описание слайда:

Minimize Per-Frame Allocations Use memory-friendly data structures, e.g. arrays, vectors Reserve memory in advance Use custom allocators Pool same-size allocations in a single block of memory to avoid overhead Use the explicit keyword to avoid hidden temporaries Avoid strings

Слайд 36

Описание слайда:

Other Tidbits Compiler settings: experiment dynamic_cast: just say no Constructors: performance killers Unused static array space: track this Loop unrolling: huge wins, sometimes Suspicious comments: watch out “Immensely slow matrix multiplication”

Слайд 37

Описание слайда:

Wrap Up Use the Right Tool for the Job The STL is Evil, Sometimes Invent Only What You Need Profile Know Thine Holy Standard Understand Your Options

Слайд 38

Описание слайда:

Call to Action: Evolve! Pass the rubber chicken Share your C++ performance mistakes with your team Mentor junior programmers So they only make new mistakes Don’t stop learning You can never know enough C++