Cache and Memory Performance

I mentioned earlier that cache latencies are higher in order to accommodate the larger caches (8MB L2 + 8MB L3) as well as the high frequency design. We turned to our old friend cachemem to measure these latencies in clocks:

Cache/Memory Latency Comparison
  L1 L2 L3 Main Memory
AMD FX-8150 (3.6GHz) 4 21 65 195
AMD Phenom II X4 975 BE (3.6GHz) 3 15 59 182
AMD Phenom II X6 1100T (3.3GHz) 3 14 55 157
Intel Core i5 2500K (3.3GHz) 4 11 25 148

Cache latencies are up significantly across the board, which is to be expected given the increase in pipeline depth as well as cache size. But is Bulldozer able to overcome the increase through higher clocks? To find out we have to convert latency in clocks to latency in nanoseconds:

Memory Latency

We disable turbo in order to get predictable clock speeds, which lets us accurately calculate memory latency in ns. The FX-8150 at 3.6GHz has a longer trip down memory lane than its predecessor, also at 3.6GHz. The higher latency caches play a role in this as they are necessary to help drive AMD's frequency up. What happens if we turn turbo on and peg the FX-8150 at 3.9GHz? Memory latency goes down. Bulldozer still isn't able to get to main memory as quickly as Sandy Bridge, but thanks to Turbo Core it's able to do so better than the outgoing Phenom II.

L3 Cache Latency

L3 access latency is effectively a wash compared to the Phenom II thanks to the higher clock speeds enabled by Turbo Core. Latencies haven't really improved though, and Bulldozer has a long way to go before it reaches Sandy Bridge access latencies.

The Impact of Bulldozer's Pipeline Windows 7 Application Performance
Comments Locked

430 Comments

View All Comments

  • Iketh - Wednesday, October 12, 2011 - link

    AMD Exec a year ago: "We about ready to release BD?"

    AMD Engineer: "Soon. At 4ghz, we're actually slower per thread and using double the power than Phenom at 3.4ghz, but we'll get there..."

    AMD Exec: /gquit
  • lyeoh - Wednesday, October 12, 2011 - link

    Bulldozer reminds me of the P4/Prescott for some reason ;).

    High clock, high watts, but not enough performance.

    Might be faster in parallelizable tasks but most people with such tasks would just buy more computers and build large clusters.
  • Iketh - Wednesday, October 12, 2011 - link

    The processors are popping up on Newegg now... the 8120 for $220 and 6100 for $190
  • vol7ron - Wednesday, October 12, 2011 - link

    Sigh... I made the mistake in buying a Prescott. Not to mention I bought an "E" batch, which ran even hotter and weren't as overclockable.
  • actionjksn - Wednesday, October 12, 2011 - link

    Yeah I had one of those hot potato's too. Back then we thought Intel was finished.
  • just4U - Thursday, October 13, 2011 - link

    ckryan, you stated you were blown away by the 2500K yes? It's odd you know.. I've owned a PII 920, PII 1055, PII 955 (tested lots of lowbie $60-80 parts from AMD to) .. also used a i7 920, i7 955 i5 2500k i7 2600k (my most recent one) and .. I am not blown away by any of them..

    Last time I was blown away by a cpu was the Q6600..(before that the A64 3200+) since then other cpu's have been better but not so much so that I'd say that it was night and day differences.
  • CeriseCogburn - Wednesday, March 21, 2012 - link

    Ok that was some enormously skilled twisting and spinning. BD is an epic failure, period. I can't envision anyone with any needs, need, or combo thereof choosing it.\
    It's so bad amd lied about it's transistor count.
    Forget it, it's an epic fail and never anything more.
  • jiffylube1024 - Wednesday, October 12, 2011 - link

    Ugh, BD is quite the disappointment. The power consumption is absolutely through the roof -- unacceptable for 32nm, really!

    With that said, I am very intrigued in the FX-4100 4-core 3.6GHz part. This should be the replacement for the Athlon II 2-4 core series, and I'm very interested to see how it does vs ~3 GHz Athlon II X2's, X3's and X4's.
  • yankeeDDL - Wednesday, October 12, 2011 - link

    Wow ...
    I'm blown away.
    I have been waiting for BD's reviews and benchmarks for months. I have waited for BD for my new rig.
    I have used AMD for the past 8 years and I am ... was convinced that it always offered, by far, the best price/performance ratio for entry-level, mid range PCs.
    I am a still a big fan of AMD ... but I have to stand corrected. BD is a POS. Longer pipelines? Didn't they learn anything from Pentium 3/4 debacle?
    A Phenom II X6 is almost always better than BD, even in power consumption. Come on: if BD had come out shortly after the Phenom I could see it as an incremental improvement, a new baseline to build upon. But it took AMD years to come out with BD ... and this is the result? Disappointing.
    I mean, betting everything on higher clock frequencies? At 4GHz? It's no wonder that Intel's IPC improvements are crunching BD: IPC is all about doing more with the same power, clock speed is all about throwing more power to do the same faster ...
    Boy. This ruined my day.
  • yankeeDDL - Wednesday, October 12, 2011 - link

    By the way, no matter how AMD slices it, I see the FX-8* as a 4-core CPU. A glorified ohene, but still a 4-core.
    If I was AMD, I would have considered a fair goal to obliterate the i5-2500 performance with the new FX-8 family, instead it comes short most of the times.
    What were they thinking?

Log in

Don't have an account? Sign up now