Recherche avancée

Sur d’autres sites (2940)

• Anatomy of an optimization : H.264 deblocking

  26 mai 2010, par Dark Shikari — H.264, assembly, development, speed, x264

  As mentioned in the previous post, H.264 has an adaptive deblocking filter. But what exactly does that mean — and more importantly, what does it mean for performance ? And how can we make it as fast as possible ? In this post I’ll try to answer these questions, particularly in relation to my recent deblocking optimizations in x264.

  H.264′s deblocking filter has two steps : strength calculation and the actual filter. The first step calculates the parameters for the second step. The filter runs on all the edges in each macroblock. That’s 4 vertical edges of length 16 pixels and 4 horizontal edges of length 16 pixels. The vertical edges are filtered first, from left to right, then the horizontal edges, from top to bottom (order matters !). The leftmost edge is the one between the current macroblock and the left macroblock, while the topmost edge is the one between the current macroblock and the top macroblock.

  Here’s the formula for the strength calculation in progressive mode. The highest strength that applies is always selected.

  If we’re on the edge between an intra macroblock and any other macroblock : Strength 4
  If we’re on an internal edge of an intra macroblock : Strength 3
  If either side of a 4-pixel-long edge has residual data : Strength 2
  If the motion vectors on opposite sides of a 4-pixel-long edge are at least a pixel apart (in either x or y direction) or the reference frames aren’t the same : Strength 1
  Otherwise : Strength 0 (no deblocking)

  These values are then thrown into a lookup table depending on the quantizer : higher quantizers have stronger deblocking. Then the actual filter is run with the appropriate parameters. Note that Strength 4 is actually a special deblocking mode that performs a much stronger filter and affects more pixels.

  One can see somewhat intuitively why these strengths are chosen. The deblocker exists to get rid of sharp edges caused by the block-based nature of H.264, and so the strength depends on what exists that might cause such sharp edges. The strength calculation is a way to use existing data from the video stream to make better decisions during the deblocking process, improving compression and quality.

  Both the strength calculation and the actual filter (not described here) are very complex if naively implemented. The latter can be SIMD’d with not too much difficulty ; no H.264 decoder can get away with reasonable performance without such a thing. But what about optimizing the strength calculation ? A quick analysis shows that this can be beneficial as well.

  Since we have to check both horizontal and vertical edges, we have to check up to 32 pairs of coefficient counts (for residual), 16 pairs of reference frame indices, and 128 motion vector values (counting x and y as separate values). This is a lot of calculation ; a naive implementation can take 500-1000 clock cycles on a modern CPU. Of course, there’s a lot of shortcuts we can take. Here’s some examples :

  • If the macroblock uses the 8×8 transform, we only need to check 2 edges in each direction instead of 4, because we don’t deblock inside of the 8×8 blocks.
  • If the macroblock is a P-skip, we only have to check the first edge in each direction, since there’s guaranteed to be no motion vector differences, reference frame differences, or residual inside of the macroblock.
  • If the macroblock has no residual at all, we can skip that check.
  • If we know the partition type of the macroblock, we can do motion vector checks only along the edges of the partitions.
  • If the effective quantizer is so low that no deblocking would be performed no matter what, don’t bother calculating the strength.

  But even all of this doesn’t save us from ourselves. We still have to iterate over a ton of edges, checking each one. Stuff like the partition-checking logic greatly complicates the code and adds overhead even as it reduces the number of checks. And in many cases decoupling the checks to add such logic will make it slower : if the checks are coupled, we can avoid doing a motion vector check if there’s residual, since Strength 2 overrides Strength 1.

  But wait. What if we could do this in SIMD, just like the actual loopfilter itself ? Sure, it seems more of a problem for C code than assembly, but there aren’t any obvious things in the way. Many years ago, Loren Merritt (pengvado) wrote the first SIMD implementation that I know of (for ffmpeg’s decoder) ; it is quite fast, so I decided to work on porting the idea to x264 to see if we could eke out a bit more speed here as well.

  Before I go over what I had to do to make this change, let me first describe how deblocking is implemented in x264. Since the filter is a loopfilter, it acts “in loop” and must be done in both the encoder and decoder — hence why x264 has it too, not just decoders. At the end of encoding one row of macroblocks, x264 goes back and deblocks the row, then performs half-pixel interpolation for use in encoding the next frame.

  We do it per-row for reasons of cache coherency : deblocking accesses a lot of pixels and a lot of code that wouldn’t otherwise be used, so it’s more efficient to do it in a single pass as opposed to deblocking each macroblock immediately after encoding. Then half-pixel interpolation can immediately re-use the resulting data.

  Now to the change. First, I modified deblocking to implement a subset of the macroblock_cache_load function : spend an extra bit of effort loading the necessary data into a data structure which is much simpler to address — as an assembly implementation would need (x264_macroblock_cache_load_deblock). Then I massively cleaned up deblocking to move all of the core strength-calculation logic into a single, small function that could be converted to assembly (deblock_strength_c). Finally, I wrote the assembly functions and worked with Loren to optimize them. Here’s the result.

  And the timings for the resulting assembly function on my Core i7, in cycles :

  deblock_strength_c : 309
  deblock_strength_mmx : 79
  deblock_strength_sse2 : 37
  deblock_strength_ssse3 : 33

  Now that is a seriously nice improvement. 33 cycles on average to perform that many comparisons–that’s absurdly low, especially considering the SIMD takes no branchy shortcuts : it always checks every single edge ! I walked over to my performance chart and happily crossed off a box.

  But I had a hunch that I could do better. Remember, as mentioned earlier, we’re reloading all that data back into our data structures in order to address it. This isn’t that slow, but takes enough time to significantly cut down on the gain of the assembly code. And worse, less than a row ago, all this data was in the correct place to be used (when we just finished encoding the macroblock) ! But if we did the deblocking right after encoding each macroblock, the cache issues would make it too slow to be worth it (yes, I tested this). So I went back to other things, a bit annoyed that I couldn’t get the full benefit of the changes.

  Then, yesterday, I was talking with Pascal, a former Xvid dev and current video hacker over at Google, about various possible x264 optimizations. He had seen my deblocking changes and we discussed that a bit as well. Then two lines hit me like a pile of bricks :

  <_skal_> tried computing the strength at least ?
  <_skal_> while it’s fresh

  Why hadn’t I thought of that ? Do the strength calculation immediately after encoding each macroblock, save the result, and then go pick it up later for the main deblocking filter. Then we can use the data right there and then for strength calculation, but we don’t have to do the whole deblock process until later.

  I went and implemented it and, after working my way through a horde of bugs, eventually got a working implementation. A big catch was that of slices : deblocking normally acts between slices even though normal encoding does not, so I had to perform extra munging to get that to work. By midday today I was able to go cross yet another box off on the performance chart. And now it’s committed.

  Sometimes chatting for 10 minutes with another developer is enough to spot the idea that your brain somehow managed to miss for nearly a straight week.

  NB : the performance chart is on a specific test clip at a specific set of settings (super fast settings) relevant to the company I work at, so it isn’t accurate nor complete for, say, default settings.

  Update : Here’s a higher resolution version of the current chart, as requested in the comments.
  

  http://git.videolan.org/?p=x264.git ;a=commit ;h=b9ec3ea27812288485329ac7771143021cd4917b

• Monster Battery Power Revisited

  28 mai 2010, par Multimedia Mike — Python, Science Projects

  So I have this new fat netbook battery and I performed an experiment to determine how long it really lasts. In my last post on the matter, it was suggested that I should rely on the information that gnome-power-manager is giving me. However, I have rarely seen GPM report more than about 2 hours of charge ; even on a full battery, it only reports 3h25m when I profiled it as lasting over 5 hours in my typical use. So I started digging to understand how GPM gets its numbers and determine if, perhaps, it’s not getting accurate data from the system.

  I started poking around /proc for the data I wanted. You can learn a lot in /proc as long as you know the right question to ask. I had to remember what the power subsystem is called — ACPI — and this led me to /proc/acpi/battery/BAT0/state which has data such as :

  present :                 yes
  capacity state :          ok
  charging state :          charged
  present rate :            unknown
  remaining capacity :      100 mAh
  present voltage :         8326 mV
  

  "Remaining capacity" rated in mAh is a little odd ; I would later determine that this should actually be expressed as a percentage (i.e., 100% charge at the time of this reading). Examining the GPM source code, it seems to determine as a function of the current CPU load (queried via /proc/stat) and the battery state queried via a facility called devicekit. I couldn’t immediately find any source code to the latter but I was able to install a utility called ’devkit-power’. Mostly, it appears to rehash data already found in the above /proc file.

  Curiously, the file /proc/acpi/battery/BAT0/info, which displays essential information about the battery, reports the design capacity of my battery as only 4400 mAh which is true for the original battery ; the new monster battery is supposed to be 10400 mAh. I can imagine that all of these data points could be conspiring to under-report my remaining battery life.

  Science project : Repeat the previous power-related science project but also parse and track the remaining capacity and present voltage fields from the battery state proc file.

  Let’s skip straight to the results (which are consistent with my last set of results in terms of longevity) :

  
  
  

  So there is definitely something strange going on with the reporting— the 4400 mAh battery reports discharge at a linear rate while the 10400 mAh battery reports precipitous dropoff after 60%.

  Another curious item is that my script broke at first when there was 20% power remaining which, as you can imagine, is a really annoying time to discover such a bug. At that point, the "time to empty" reported by devkit-power jumped from 0 seconds to 20 hours (the first state change observed for that field).

  Here’s my script, this time elevated from Bash script to Python. It requires xdotool and devkit-power to be installed (both should be available in the package manager for a distro).
  

  PLAIN TEXT

  PYTHON :

  1. # !/usr/bin/python
  2.  
  3. import commands
  4. import random
  5. import sys
  6. import time
  7.  
  8. XDOTOOL = "/usr/bin/xdotool"
  9. BATTERY_STATE = "/proc/acpi/battery/BAT0/state"
  10. DEVKIT_POWER = "/usr/bin/devkit-power -i /org/freedesktop/DeviceKit/Power/devices/battery_BAT0"
  11.  
  12. print "count, unixtime, proc_remaining_capacity, proc_present_voltage, devkit_percentage, devkit_voltage"
  13.  
  14. count = 0
  15. while 1 :
  16.   commands.getstatusoutput("%s mousemove %d %d" % (XDOTOOL, random.randrange(0,800), random.randrange(0, 480)))
  17.   battery_state = open(BATTERY_STATE).read().splitlines()
  18.   for line in battery_state :
  19.     if line.startswith("remaining capacity :") :
  20.       proc_remaining_capacity = int(line.lstrip("remaining capacity : ").rstrip("mAh"))
  21.     elif line.startswith("present voltage :") :
  22.       proc_present_voltage = int(line.lstrip("present voltage : ").rstrip("mV"))
  23.   devkit_state = commands.getoutput(DEVKIT_POWER).splitlines()
  24.   for line in devkit_state :
  25.     line = line.strip()
  26.     if line.startswith("percentage :") :
  27.       devkit_percentage = int(line.lstrip("percentage :").rstrip(’\%’))
  28.     elif line.startswith("voltage :") :
  29.       devkit_voltage = float(line.lstrip("voltage :").rstrip(’V’)) * 1000
  30.   print "%d, %d, %d, %d, %d, %d" % (count, time.time(), proc_remaining_capacity, proc_present_voltage, devkit_percentage, devkit_voltage)
  31.   sys.stdout.flush()
  32.   time.sleep(60)
  33.   count += 1

• Revision 35682 : maj invalideur

  28 février 2010, par brunobergot@… — Log

  maj invalideur