The Impact of Bulldozer's Pipeline

With a new branch prediction architecture and an unknown, but presumably significantly deeper pipline, I was eager to find out just how much of a burden AMD's quest for frequency had placed on Bulldozer. To do so I turned to the trusty N-Queens solver, now baked into the AIDA64 benchmark suite.

The N-Queens problem is simple. On an N x N chessboard, how do you place N queens so they cannot attack one another? Solving the problem is incredibly branch intensive, and as a result it serves as a great measure of the impact of a deeper pipeline.

The AIDA64 implementation of the N-Queens algorithm is heavily threaded, but I wanted to first get a look at single-core performance so I disabled all but a single integer/fp core on Bulldozer, as well as the competing processors. I also looked at constant frequency as well as turbo enabled speeds:

Single Core Branch Predictor Performance—AIDA64 Queens Benchmark

Unfortunately things don't look good. Even with turbo enabled, the 3.6GHz Bulldozer part needs another 25% higher frequency to equal a 3.6GHz Phenom II X4. Even a 3.3GHz Phenom II X6 does better here. Without being fully aware of the optimizations at work in AIDA64 I wouldn't put too much focus on Sandy Bridge's performance here, but Intel is widely known for focusing on branch prediction performance.

If we let the N-Queens benchmark scale to all available threads, the performance issues are easily masked by throwing more threads at the problem:

SMP Branch Predictor Performance—AIDA64 Queens Benchmark

However it is quite clear that for single or lightly threaded operations that are branch heavy, Bulldozer will be in for a fight.

Power Management and Real Turbo Core Cache and Memory Performance
Comments Locked

430 Comments

View All Comments

  • silverblue - Monday, October 17, 2011 - link

    The reason for mentioning the 2600 is because that's the only comparison we have for the moment. I don't expect Valencia to use as much power as Zambezi even on a clock-for-clock basis.
  • Pipperox - Monday, October 17, 2011 - link

    That has to be seen.
    Bulldozer's HUGE 16MB of total cache surely were not put there for desktop workloads.
  • xineis - Sunday, October 16, 2011 - link

    I am quite disappointed with the overall performance. And specially with gaming performance.

    But what about that rumored patch fix for the BDZ? Has any of you guys at AnandTech heard of that?
  • silverblue - Sunday, October 16, 2011 - link

    You mean this one?

    http://quinetiam.com/?p=2356

    Interesting if it's true, however it's doubtful to help across the board (pun not intended). Still, very good news for those doing rendering/encoding.
  • richaron - Sunday, October 16, 2011 - link

    I never want to see a link to that dickbag again..
    His blogs about WW3, bigfoot, & 2012 should be enough to give you an idea. Also, the dudes at Kubuntu basically do repackaging & KDE integration. They don't touch hardware.
    No doubt there can be significant gains through software. But I would rather stab myself through the eyelid than read anything more by that 'person'.
  • xineis - Sunday, October 16, 2011 - link

    Yeah, that one. I know that there is a very slim chance of that 40% boost ever happening, but the idea behind the "patch" is actually plausible.

    I've seen lots of people explaning how would it work and it looks somewaht legit.

    Granted that 40% is too much, but if the right tweaking, I'd be more than happy with 20% to 30% increase in performance!
  • richaron - Monday, October 17, 2011 - link

    everyone would be happy with 20-30%. It doesn't seem realistic to me... but i dropped out of computer engineering :p

    From what I've seen, I would guess about 5% [+/- 5%]
  • silverblue - Monday, October 17, 2011 - link

    Turns out he was one giant troll, or seemingly so. Now, he's pointing the finger at...

    "Problem solved, it’s just a thermal protection issue, people have been pushing voltages too high. Maybe there’s some variance in mainboard chipsets, but some overclockers and hitting really good numbers."

    Really? After all that hoo-har about registry patches, BIOS flashes and the like, we're now blaming thermal protection? I'm taking this with a litre of Dead Sea water.
  • silverblue - Monday, October 17, 2011 - link

    Hey, there has to be somebody somewhere that we can all laugh at. ;)

    That said, he does sound a bit... odd.
  • Romulous - Monday, October 17, 2011 - link

    perhaps Bulldozer.Next will actually work ;)

Log in

Don't have an account? Sign up now