This site may earn chapter commissions from the links on this page. Terms of use.

Zen-Feature

Last calendar week, the commencement purported benchmarks of AMD's upcoming "Zen" microprocessor leaked, giving us an early glimpse of how the chip performs vis-à-vis diverse Intel products. While the leaks haven't been confirmed as genuine, it was interesting to encounter the conversation around Zen and AMD'south expected operation level. Then, at an event on Midweek, AMD showed some demos of Zen running clock-for-clock against Intel'due south 6900K. While these demos were not performed at full speed, they offered some evidence that AMD could lucifer Intel clock-for-clock in sure workloads — and while such demos are always cherry-picked to favor ane's ain company, they gave some boosted context on what Zen can do.

What I desire to talk about here has less to practise with AMD's demo or specifics of the architecture and more to do with what AMD needs to evangelize to exist considered a success. One thing I'll say upwardly front: I have no secret sources, no hidden informants, and no leaks to credit for what I'm about to write. Agree or disagree, these are my own thoughts and observations.

AMD-Zen-02

AMD claims that Zen will have 40% meliorate IPC than its previous Excavator compages and it's shown demos illustrating how Zen tin compete clock-for-clock against the 6900K in some workloads. This is both a solid indication of where the bit currently is and a generalized prediction of where information technology could end upwards. The details and specifics will depend on a number of factors including terminal clock speeds and application-level optimization.

What does forty% more IPC hateful, anyway?

IPC stands for instructions per clock or instructions per cycle (the two terms are synonymous). It's a metric of CPU efficiency — the higher a CPU'due south IPC, the more work that CPU tin perform in a given amount of time.

In that location'south typically an inverse human relationship betwixt a CPU's clock speed and its IPC. While IPC has always been used to measure CPU performance, the term became common in enthusiast circles when Intel debuted the original Pentium 4. While the P4 was clocked much higher than its Pentium 3 predecessor, the amount of work it performed per clock bike was significantly lower. The cyberspace result of this was that the Pentium 4 i.5GHz oftentimes struggled to outperform the slower Pentium three 1GHz or its Athlon K7 analogue from AMD.

The most common way to parse AMD's argument is as follows: If you took an Excavator CPU and a Zen CPU and ran them at the aforementioned clock speed, the Zen CPU should be 40% faster, on average, and then the AMD CPU information technology replaces. But even here, there have been questions: Does that xl% IPC improvement include the affect of simultaneous multi-threading (what Intel calls Hyper-Threading) or not? Is that based on two chips clocked at the same frequency or not? Finally and most importantly, which tests were used to derive that figure? IPC isn't a constant — it fluctuates a nifty bargain based on the beliefs of the target application.

Maxwell-Performance

The benchmark above is from tests I ran when Bulldozer launched well-nigh 5 years agone. Despite losing to AMD'due south older 6-core X6 1100T in multiple benchmarks, the FX-8150 was really significantly faster in Maxwell Render — one of the few tests it unilaterally won. This kind of variance is normal, which is why it'southward so difficult to draw conclusions nigh relative chip performance based on any single metric. AMD's claimed 40% IPC uplift should be treated like a general or average prediction and not the guaranteed result of whatsoever single exam.

How will Zen compare with Intel CPUs?

The real question, of course, isn't whether Zen will meliorate on Carrizo, just whether it'll requite AMD a CPU that tin compete with Intel. The realistic reply is "It'll depend on where you look and what you're looking for."

Here'southward what I mean by that. Compare AMD's current top-terminate Piledriver, the FX-9590, against Intel'due south Core i7-6700K using Anandtech's Bench tool. I've snipped a section of tests to include beneath, but yous can view the total comparison here. In the graph below, blueish is for the FX-9590, while the orange-ish bar is Intel'south Core i7-6700K.

Assume for a moment that these FX-9590 results actually reflected Excavator performance instead of Piledriver. Now, assume they were 40% faster than they actually are. Would this theoretical top-end chip offer competitive performance with Intel? In some tests, it absolutely would. In other tests, the gap between AMD and Intel is large enough that fifty-fifty a xl% improvement isn't plenty to let AMD to take hold of its competitor. Those of y'all who want to perform this comparison using bodily data from Excavator tin refer to this AT review of the scrap and exercise the math from at that place. Either way, the point holds — in that location are tests where a xl% improvement would absolutely be enough to allow AMD to catch Intel and tests where it wouldn't be. When nosotros say "It's going to depend," information technology'southward non a dodge or an alibi, it's a fact. AMD's IDF showcase gave some additional information but non enough to modify this basic situation.

How do y'all swallow an elephant?

1 of the common arguments raised in the Ashes of the Singularity benchmark thread was that Zen's comparatively weak performance compared with the Core i7-6900K is proof that AMD's eight-core chip won't be able to compete confronting Intel. The uncomplicated fact that these test results were run on an ES chip of unknown stepping and vintage make that an extremely premature determination. Worse, it's a determination founded in an incorrect premise — specifically, that AMD has to friction match Intel's top-end operation in order to compete at all. While AMD's recent demos should assuage some of these fears, the early unveil wasn't meant to exist a complete overview of every aspect of the flake'southward performance.

Correct at present, AMD's FX-8350 is a $160 CPU, while the summit-cease FX-9590 retails for $229. Intel's quad-core Core i5 desktop processors start at $185 for older Haswell parts and $190 for Skylake. The cheapest Core i7 yous can buy starts at $295 for Haswell and $305 for Skylake (all prices from NewEgg). This means the FX-8350 is priced confronting a high-end Skylake Core i3 (the Core i3-6320) while the FX-9590 faces off confronting the Cadre i5-6600 — a comparatively tiny chip with an integrated GPU and one-third the TDP.

In the 15 years I've been reviewing CPUs, the gap between AMD and Intel has never been larger than it is today. That'southward role of why Zen is so important, only it's also why it'south important to keep perspective on what kind of functioning improvement AMD can reasonably deliver in a unmarried product bike.

Based on AMD's unmarried 40% IPC effigy, it's extremely unlikely that the company volition deliver a CPU that can match Intel's functioning at every particular and at every price point — and it doesn't need to. A 40% IPC boost would allow AMD to challenge Intel'due south Core i3/i5/i7 line-up by leveraging larger core counts and its ain SMT implementation far more than effectively than Piledriver ever did. The drastically reduced TDP (Zen has reportedly targeted 95W at the high finish) will let it to compete much more aggressively on power consumption. Higher CPU efficiency means that AMD won't need to rely on high clocks to hit performance targets, making it easier to push into laptops when Zen-based APUs come to market. Trade-offs between core count and clock speed also mean AMD can probably offering lower core counts at college clocks, the same fashion that Intel does.

Information technology's easy to forget, but AMD didn't launch K7 in 1999 and seize twenty% of the server market 12 months later. K7 was competitive with the Pentium 3, just not necessarily faster — peculiarly since Intel's Coppermine P3's ran a full-speed L2 cache while the slot-based K7 and K7.five used a half-speed or one/3 speed L2. Socket A and Thunderbird airtight this gap in June 2000, but AMD didn't enter the server market until 2001. It didn't see serious success in the server market until 2003, when K8 gave it the legs it needed to become head-to-caput with Intel's high-terminate Xeons.

Right now, Intel's least expensive eight-core processor is the Core i7-5960X, at $1,015. AMD doesn't demand to hit Intel in the $1,000 CPU market to drastically improve its own fortunes or competitive standing.

Conclusion

Zen will be judged a failure or a success based on how well it performs today and on how well it positions AMD to deliver future improvements. It may non going to exist potent enough to go toe-to-toe with Skylake across every SKU and toll betoken, merely that's not what it needs to exercise. AMD needs Zen to perform well at low power so information technology can be paired with an APU and slipped into notebooks. It needs Zen to be more efficient than Excavator then Sunnyvale isn't left pricing eight-cadre chips against Intel's Core i3. And it needs Zen to be strong enough that investors and enthusiasts see the core as a futurity to be built upon, rather than an anchor around the company's cervix.

I genuinely don't know how well AMD will evangelize on these goals. But those are the criteria I expect the chip to need to fulfill and the metrics by which I'll judge its overall position. I'll have more to say most Zen, including a deep swoop into its architecture, when I render adjacent week.