20200722, 06:30  #1 
Sep 2006
The Netherlands
1273_{8} Posts 
64x64 integer multiplication GCC
Good Morning!
Looking for way of multiplying fast in GCC 64x 64 bits unsigned 64 bits integers and obtaining both highres as well as lowres result. Was busy reviving CUDA sieving code i wrote in 2016 and to verify the results on the CPU my oldie code there for GCC is duck slow. Under windows it's fast enough there as one has intrinsics: reslo = _umul128(x,y,&reshi); However what i have for under linux where my GPU runs under i have only something duck slow for GCC. Right now under GCC i'm using the duckslow next thing (typing over by hand): uint64 mul_limbs(uint64 *reslo,uint64 x,uint64 y) { __uint128_t prod; prod = (__uint128_t) x * y; *reslo = (uint64)prod; return(prod>>64); } This duckslow  anyone know faster solution  and i just need multiply a single x * y for the BSGS algorithm verification. Kind Regards, Vincent 
20200722, 13:25  #2 
"Ben"
Feb 2007
2^{5}×3×5×7 Posts 
Try:
Code:
__inline uint64_t mul64(uint64_t x, uint64_t y, uint64_t* hi) { __asm__( "mulq %3 \n\t" : "=&a"(x), "=&d"(y) : "0"(x), "1"(y) : "cc" ); *hi = y; return x; } Code:
__inline uint64_t mulx64(uint64_t x, uint64_t y, uint64_t* hi) { __asm__( "mulx %3, %0, %1 \n\t" : "=&d"(x), "=&a"(y) : "0"(x), "1"(y) ); *hi = y; return x; } Last fiddled with by bsquared on 20200722 at 13:47 Reason: slight change to mul64 
20200722, 14:14  #3 
Tribal Bullet
Oct 2004
2·3·19·31 Posts 
As a slight modification, the following macro (from an old version of GMP's longlong.h) can avoid a potential trip through memory:
Code:
#define umul_ppmm(w1, w0, u, v) \ __asm__ ("mulq %3" \ : "=a" (w0), "=d" (w1) \ : "%0" ((UDItype)(u)), "rm" ((UDItype)(v))) 
20200722, 14:56  #4  
"Ben"
Feb 2007
2^{5}·3·5·7 Posts 
Quote:
However, for this simple benchmark loop: Code:
for (i = 0; i < 1000000000; i++) { c = mulx64(a, b, &d); a = c; b = d; } Code:
mulx took 0.2715 sec mul took 0.2714 sec mul_limbs took 0.2715 sec umul_ppmm took 0.2714 sec 

20200722, 15:08  #5 
"Oliver"
Sep 2017
Porta Westfalica, DE
406_{10} Posts 
Is GCC optimizing it to the same target assembly code?
Last fiddled with by kruoli on 20200722 at 15:08 Reason: Spelling. 
20200722, 15:22  #6 
Sep 2006
The Netherlands
3·233 Posts 
i wrote that mul_limbs code like 15 years ago. maybe GCC got a lot more clever lately :)
Back then you can bet i cried it out loud like a wounded Bear that GCC was duckslow optimizing it back then in my NTT code versus oldie visual c++ compiler making fast clean code with its intrinsics (back then not so old). I don't know which GCC version gets called by Nvidia's Cuda compiler driver. nvcc version gives 'built on sun_sep__4_22:14:01_cdt_2016 gcc version gives 4.9.2 Let's see whether NVCC knows the S option to print assembler. If NVCC is not clever enough and this GCC version is  i can still consider making a library out of the math functions and just simply link it. 
20200722, 15:48  #7 
"Ben"
Feb 2007
6440_{8} Posts 

20200722, 23:04  #8  
∂^{2}ω=0
Sep 2002
República de California
2^{2}·3·5·193 Posts 
Quote:
Last fiddled with by ewmayer on 20200722 at 23:06 

20200723, 02:42  #9 
"Ben"
Feb 2007
2^{5}×3×5×7 Posts 
Agreed, it's not a good benchmark. The OP asked for a way to multiply 64bit unsigned ints and he now has several things to try. I assumed he will use them in his application to figure out which is best... the loop was just a first cut to see if there were glaring differences.

20200723, 08:24  #10 
Undefined
"The unspeakable one"
Jun 2006
My evil lair
1011101101010_{2} Posts 
I am still amazed at how much C code it takes just to coax it into using a single native instruction.
You have to throw out all the "portability" that C is supposed to provide. And you have to provide a lot of ugly boilerplate lines to coerce it into assembling just one line. And you have to write it the worst possible syntax: AT&T. 
20200723, 08:34  #11  
Bamboozled!
"𒉺𒌌𒇷𒆷𒀭"
May 2003
Down not across
2900_{16} Posts 
Quote:
C is an old language designed to be efficient on a computer, the PDP11, which was discontinued years before its manufacturer, DEC, went out of business  itself many years ago. AFAICT essentially all other languages (assembly excepted of course) make it difficult to exploit all the instructions provided by any CISC architecture. 

Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Even faster integer multiplication  paulunderwood  Computer Science & Computational Number Theory  17  20200521 19:51 
multiplication and logarithm  bhelmes  Math  4  20161006 13:33 
k*b^n+/c where b is an integer greater than 2 and c is an integer from 1 to b1  jasong  Miscellaneous Math  5  20160424 03:40 
New Multiplication Algorithm  vector  Miscellaneous Math  10  20071220 18:16 
Multiplication Tendency  clowns789  Miscellaneous Math  5  20050311 00:23 