Recherche avancée

Sur d’autres sites (8219)

• Summer of Code (in space)

  25 juillet 2011, par Dark Shikari — development, SOCIS

  There’s apparently another Summer of Code in town. x264 has been accepted into the ESA Summer of Code in Space. Just like Google Summer of Code, work on x264 over the summer and get paid ! Watch out though ; only some countries are allowed, so check first if you’re allowed to participate. The application deadline is July 27, 11AM (UTC) ; sorry for the short notice this time around !

• Recapping WebM’s First Week

  25 mai 2010, par noreply@blogger.com (John Luther) — webm, vp8, vorbis

  The WebM project launched last Wednesday with broad industry backing (watch video of the announcement). The list of supporters keeps growing with new additions such as the popular VLC media player, Miro Video Converter, HeyWatch cloud encoding platform, and videantis programmable processor platform. We’re also happy to see that future versions of IE will support playback of VP8 when the user has installed the codec.

  Our announcement sparked discussions in the community around the design and quality of our developer release. We’ve done extensive testing of VP8 and know that the codec can match or exceed the quality of other leading codecs. Starting this week, the engineers behind WebM will post frequently to this blog with details on how to make optimal use of its VP8 video codec and Vorbis audio codec. We are confident that the open development model will bring additional improvements that will further optimize WebM. In fact, the power of open development is already visible, with developers submitting patches and the folks at Flumotion enabling live streaming support in their product just three days after the project was launched.

  Keep an eye on this blog for regular updates on the adoption and development of WebM. To participate in the conversation or to ask questions of the WebM team, please join our discussion group.

  John Luther
  Product Manager, Google

  

• Bit-field badness

  30 janvier 2010, par Mans — Compilers, Optimisation

  Consider the following C code which is based on an real-world situation.

  struct bf1_31 
      unsigned a:1 ;
      unsigned b:31 ;
   ;
  void func(struct bf1_31 *p, int n, int a)
  
  
  
      int i = 0 ;
  
      do 
  
          if (p[i].a)
  
              p[i].b += a ;
  
       while (++i < n) ;
  
  
  

  How would we best write this in ARM assembler ? This is how I would do it :
  

  func :
          ldr     r3,  [r0], #4
          tst     r3,  #1
          add     r3,  r3,  r2,  lsl #1
          strne   r3,  [r0, #-4]
          subs    r1,  r1,  #1
          bgt     func
          bx      lr
  

  The add instruction is unconditional to avoid a dependency on the comparison. Unrolling the loop would mask the latency of the ldr instruction as well, but that is outside the scope of this experiment.

  Now compile this code with gcc -march=armv5te -O3 and watch in horror :

  func :
          push    r4
          mov     ip, #0
          mov     r4, r2
  loop :
          ldrb    r3, [r0]
          add     ip, ip, #1
          tst     r3, #1
          ldrne   r3, [r0]
          andne   r2, r3, #1
          addne   r3, r4, r3, lsr #1
          orrne   r2, r2, r3, lsl #1
          strne   r2, [r0]
          cmp     ip, r1
          add     r0, r0, #4
          blt     loop
          pop     r4
          bx      lr
  

  This is nothing short of awful :

  • The same value is loaded from memory twice.
  • A complicated mask/shift/or operation is used where a simple shifted add would suffice.
  • Write-back addressing is not used.
  • The loop control counts up and compares instead of counting down.
  • Useless mov in the prologue ; swapping the roles or r2 and r4 would avoid this.
  • Using lr in place of r4 would allow the return to be done with pop {pc}, saving one instruction (ignoring for the moment that no callee-saved registers are needed at all).

  Even for this trivial function the gcc-generated code is more than twice the optimal size and slower by approximately the same factor.

  The main issue I wanted to illustrate is the poor handling of bit-fields by gcc. When accessing bitfields from memory, gcc issues a separate load for each field even when they are contained in the same aligned memory word. Although each load after the first will most likely hit L1 cache, this is still bad for several reasons :

  • Loads have typically two or three cycles result latency compared to one cycle for data processing instructions. Any bit-field can be extracted from a register with two shifts, and on ARM the second of these can generally be achieved using a shifted second operand to a following instruction. The ARMv6T2 instruction set also adds the SBFX and UBFX instructions for extracting any signed or unsigned bit-field in one cycle.
  • Most CPUs have more data processing units than load/store units. It is thus more likely for an ALU instruction than a load/store to issue without delay on a superscalar processor.
  • Redundant memory accesses can trigger early flushing of store buffers rendering these less efficient.

  No gcc bashing is complete without a comparison with another compiler, so without further ado, here is the ARM RVCT output (armcc --cpu 5te -O3) :

  func :
          mov     r3, #0
          push    r4, lr
  loop :
          ldr     ip, [r0, r3, lsl #2]
          tst     ip, #1
          addne   ip, ip, r2, lsl #1
          strne   ip, [r0, r3, lsl #2]
          add     r3, r3, #1
          cmp     r3, r1
          blt     loop
          pop     r4, pc
  

  This is much better, the core loop using only one instruction more than my version. The loop control is counting up, but at least this register is reused as offset for the memory accesses. More remarkable is the push/pop of two registers that are never used. I had not expected to see this from RVCT.

  Even the best compilers are still no match for a human.