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
POST A COMMENT

428 Comments

View All Comments

  • Mishera - Monday, October 17, 2011 - link

    Good points TekDemon. But I'll add that from what I understand, the GPU might be capable of processing huge amounts of graphic information, but might have to wait for the CPU to process certain information before it's able to continue, hence some games going only so high in graphic tests no matter what kind of GPU is put in.

    Like he said, buying a good CPU will last longer than spending that money on a really good GPU. I personally try to build a balanced system since by the time I upgrade it's a pretty big jump on all ends.
    Reply
  • Snorkels - Wednesday, October 12, 2011 - link

    This and other benchmark tests are BOGUS. You are comparing apples to oranges. LiarMarks..

    This test shows non-optimized code for AMD vs optimized code for the Intel CPU.

    It does not show the actual performance of the Bulldozer CPU.

    Most software companies compile their software using Intels compiler, which creates crappy and unefficient codepaths for AMD processors.

    Compile with Open64 compiler and you get a totally different result.
    Reply
  • actionjksn - Wednesday, October 12, 2011 - link

    @Snorkels If most software company's are using an Intel compiler, why would you want an AMD processor that can't utilize it properly. Reply
  • g101 - Wednesday, October 12, 2011 - link

    Well, I'm happy that gamer children cannot understand the point of this architecture. You obviously have no concept of the architectural advantages, since it's not designed for game-playing children or completely unoptimized synthetic benchmarks.

    Bulldozer optimized and future-proofed for professional software, rather than entertainment software for children.
    Reply
  • AssBall - Wednesday, October 12, 2011 - link

    "obviously have no concept of the architectural advantages"

    Enlighten us then, oh wise one.
    Reply
  • guyjones - Sunday, October 16, 2011 - link

    Who exactly is the child here? Your infantile comment conveniently ignores the fact that AMD has made gigantic marketing pushes that are clearly directed at the gamer community, not to mention gaming-related sponsorship activities and marketing tie-ins. So, on the contrary, the company has made very visible and consciously-directed efforts to appeal to gamers with its products. It is totally unreasonable to now posit that BD is not directed at least in part toward that market segment. Reply
  • Will Robinson - Wednesday, October 12, 2011 - link

    FailDozer....pretty limp performance numbers.
    Intel still rules.
    Reply
  • etrigan420 - Wednesday, October 12, 2011 - link

    What an unfortunate series of events...maybe my e8400 will hold out a little longer... Reply
  • Malih - Wednesday, October 12, 2011 - link

    before reading (and i've read all other review sites -and disappointed at AMD-, just dying to see your view on the matter) thanks for the review as always Reply
  • xorbit - Wednesday, October 12, 2011 - link

    You are not measuring "branch prediction" performance. You are measuring misspeculation penalty (due to longer pipeline or other reasons). Nothing can "predict" random data-dependent branches. Reply

Log in

Don't have an account? Sign up now