|
|
MMX register to SSE register
Hello,
I need to move a value from a MMX to a SSE register.
I found two options.
The first one is using the instructions movq2dq %mm0, %xmm0.
The second one is moving the value from %mm0 to memory and then from memory to %xmm0. For example:
movq %mm0, -16(%rsp)
movq -16(%rsp), %xmm0
The last one is generated by gcc.
I was doing some test and movq2dq is significantly slower that the second option and I was wondering why, because then... what is the usefulness of the instruction?
Any idea?
Thank you much.
|
|
0
|
|
|
|
Reply
|
 _m256
|
4/26/2011 10:01:13 AM |
|
|
0 Replies
2708 Views
(page loaded in 0.036 seconds)
Similiar Articles: MMX reverse byte order instr. - comp.lang.asm.x86Hi I have read in the internet that there is MMX instruction that reverses byte order in MMX register, however i can't find this instruction name on... How to find out which instruction sets (mmx/sse/etc.) are used ...... Register: How to find out which instruction sets (mmx/sse/etc.) are used? Follow SSE Programming - comp.lang.asm.x86Emits the Streaming SIMD Extensions 4 (SSE4) instruction popcnt. ... SSE3 and SSE4 Streaming ... ... Likewise since the cost of moving data to and from the MMX register file from ... How to perform integer div with SIMD? - comp.lang.asm.x86 ...Hi, is there a way to make use of MMX/SSE to do calculations like 'a mod b' where ... integer div with SIMD? - comp.lang.asm.x86 ... How to get integer value for register ... fpu code optimisation request - comp.lang.asm.x86Many instructions (e.g. shufps/shufpd/pshufd) are only distinct because the FP register file, which is shared between x87, MMX, and SSE, keeps track of the type of data ... How to cast using MSVC++ Intrinsics - comp.lang.asm.x86MSVC++ serialized the load/store using one register > ignoring a more efficient ... How to find out which instruction sets (mmx/sse/etc.) are used ... How to cast using ... 128-bit MMX versus 32-bit memory copy - comp.lang.asm.x86 ...... Register ... I figured I'd get a significant speedup if I used MMX ... you talking about MMX (movq), which is 64-bit, or SSE ... Fitting data and then accessing sse, rsquare, dfe, adjrsquare and ...After fitting these I want to access the sse, rquare, dfe, adjrsquare and rmse ... Register Re: Using SSE 128 bit movs From One Memory Location To Another ...... it runs 10 times faster than either my= inlined SSE ... 128-bit MMX versus 32-bit memory copy - comp.lang.asm ... byte from ... to-memory, then use MOV or XCHG for register ... Fast bit-reverse on an x86? - comp.dspFor IA32 there is BSWAP to swap the bytes in a 32 bit register, and XCHG to swap ... > > [...] > > MMX/SSE instructions > > Yes, like (Warning! untested code follows): > > movdqa ... Streaming SIMD Extensions - Wikipedia, the free encyclopediaStreaming SIMD Extensions ... register set (the XMM registers), and it adds a few integer instructions that work on MMX registers. SSE was ... is a registered ... MMX (instruction set) - Wikipedia, the free encyclopedia... of a full 80-bit FPU register). The main usage of the MMX ... MMX (1996) · 3DNow! (1998) Streaming SIMD Extensions (SSE) (1999) · SSE2 (2001 ... Wikipedia® is a registered ... 7/23/2012 6:18:11 PM
|
|
|
|
|
|
|
|
|
Search |
Post Question |
Groups |
Stream |
About |
Register
10 followers. 