Core: It’s all in the Prefetch

In a simple CPU design, instructions are decoded in the core and data is fetched from the caches. In a perfect world, such as the Mill architecture, the data and instructions are ready to go in the lowest level cache at all times. This allows for the lowest latency and removes a potential bottleneck. Real life is not that rosy, and it all comes down to how the core can predict what data it needs and has enough time to drag it down to the lowest level of cache it can before it is needed. Ideally it needs to predict the correct data, and not interfere with memory sensitive programs. This is Prefetch.

The Core microarchitecture added multiple prefetchers in the design, as well as improving the prefetch algorithms, to something not seen before on a consumer core. For each core there are two data and one instruction prefetchers, plus another couple for the L2 cache. That’s a total of eight for a dual core CPU, with instructions not to interfere with ‘on-demand’ bandwidth from running software.

One other element to the prefetch is tag lookup for cache indexing. Data prefetchers do this, as well as running software, so in order to avoid a higher latency for the running program, the data prefetch uses the store port to do this. As a general rule (at least at the time), loads happen twice as often as stores, meaning that the store port is generally more ‘free’ to be used for tag lookup by the prefetchers. Stores aren’t critical for most performance metrics, unless the system can’t process stores quickly enough that it backs up the pipeline, but in most cases the rest of the core will be doing things regardless. The cache/memory sub-system is in control for committing the store through the caches, so as long as this happens eventually the process works out.

Core: More Cache Please

Without having access to a low latency data and instruction store, having a fast core is almost worthless. The most expensive SRAMs sit closest to the execution ports, but are also the smallest due to physical design limitations. As a result, we get a nested cache system where the data you need should be in the lowest level possible, and accesses to higher levels of cache are slightly further away. Any time spent waiting for data to complete a CPU instruction is time lost without an appropriate way of dealing with this, so large fast caches are ideal. The Core design, over the previous Netburst family but also over AMD’s K8 ‘Hammer’ microarchitecture, tried to swat a fly with a Buick.

Core gave a 4 MB Level 2 cache between two cores, with a 12-14 cycle access time. This allows each core to use more than 2MB of L2 if needed, something Presler did not allow. Each core also has a 3-cycle 32KB instruction + 32KB data cache, compared to the super small Netburst, and also supports 256 entries in the L1 data TLB, compared to 8. Both the L1 and L2 are accessible by a 256-bit interface, giving good bandwidth to the core.

Note that AMD’s K8 still has a few advantages over Core. The 2-way 64KB L1 caches on AMD’s K8 have a slightly better hit rate to the 8-way 32KB L1 caches on Core, with a similar latency. AMD’s K8 also used an on-die memory controller, lowering memory latency significantly, despite the faster FSB of Intel Core (relative to Netburst) giving a lower latency to Core. As stated in our microarchitecture overview at the time, Athlon 64 X2s memory advantage had gotten smaller, but a key element to the story is that these advantages were negated by other memory sub-system metrics, such as prefetching. Measured by ScienceMark, the Core microarchitecture’s L1 cache delivers 2x bandwidth, and the L2 cache is about 2.5x faster, than the Athlon one.

Ten Year Anniversary of Core 2 Duo and Conroe Core: Decoding, and Two Goes Into One
POST A COMMENT

159 Comments

View All Comments

  • e1jones - Wednesday, July 27, 2016 - link

    My E8400 is still my daily driver, 4x 2gb and an SSD swapped in later as the boot drive. Still runs great, except it tends to get bogged down by the TrustedInstaller and the Firefox memory leaks. Reply
  • rarson - Friday, August 5, 2016 - link

    I've got an E8600 in an Abit IP35 Pro motherboard. I was having a hard time finding DDR2-1066 last I looked, so I settled for 800. With an SSD and 7870, it's surprising how well it still games. I don't think I'll upgrade the GPU again just due to the fact that I'm limited to PCI-e 2. Reply
  • FourEyedGeek - Monday, August 8, 2016 - link

    You could get a higher end GPU and still benefit from increased performance, then get a new CPU motherboard combo when you want too. Reply
  • BrokenCrayons - Wednesday, July 27, 2016 - link

    I just upgraded out of a Q6600 and 4GB DDR2 about 2 months ago and I admit that I was still kicking around the idea of leaving it alone as I was pulling the motherboard out of the case. I replaced it with a cheap AMD 860k and 16GB DDR3 which really hasn't done a lot to improve the system's performance. In retrospect, I think I could realistically have squeezed another couple of years out of it, but the motherboard's NIC was iffy and I really wanted reliable ethernet.

    As for laptops, I've got a couple C2Ds kicking around that are perfectly adequate (T2310 & P8400) for daily use. I really can't see any point in replacing them just yet. Core was a good design through all its iterations.
    Reply
  • Beany2013 - Wednesday, July 27, 2016 - link

    I like your style - rather than drop $100 on a midlevel intel NIC, you replace an entire platform.

    I strongly approve of these economics :-)
    Reply
  • Michael Bay - Thursday, July 28, 2016 - link

    USB3 is kind of nice. Reply
  • BrokenCrayons - Thursday, July 28, 2016 - link

    Well the NIC wasn't the only reason, but it was the last in a series of others that I was already coping with that tipped the scales. The upgrade was under $200 for the board, processor and memory so it really boiled down to one weekend dinner out to a mid-range restaurant. It was worth it for more reliable Steam streaming and fewer VNC disconnects as that wired ethernet port is the only means by which I regularly interact with my desktop since it has no monitor and is crammed into a corner in my utility room. Reply
  • artk2219 - Friday, July 29, 2016 - link

    Why didn't you go for an FX if you dont mind me asking? You liked the FM2+ platform a bit better? Reply
  • BrokenCrayons - Friday, July 29, 2016 - link

    Actually, I didn't give much of anything in the system a very close look before buying. I admittedly did about twenty minutes of research to make sure the 860k and the bottom feeder motherboard I'd picked would play nicely together before making a purchase. So the CPU & motherboard pair were the result of laziness and apathy rather than a preference for FM2+. Reply
  • artk2219 - Monday, August 1, 2016 - link

    Ah ok gotcha, I just wanted to share that if you had a microcenter near you they sell FX 8320E's bundled with motherboards for 125 to 170 depending on which board you want to use. That can be quite the steal and a great base for a new cheap system once you bump the clocks on the 8320E. Reply

Log in

Don't have an account? Sign up now