Buying an 8-core processor was a wallet ripping affair prior to the arrival of Ryzen. In absence of competition, the Intel Broadwell-E Core i7-6900K was ridiculously overpriced at $1,050. Intel has had to make changes to its HEDT platform by releasing the Core i7-7820X in response to the 8-core/16-thread Ryzen 7 1700.
However, I'm not quite sure Intel understands how competition works. Whereas AMD's solution launched with an MSRP of $330 and can be readily purchased online for as little as $290 today, Intel is still asking $600 for the i7-7820X. Worse still, the cheaper Ryzen 7 1700 can be happily paired with a $100 motherboard, while the pricier Core i7-7820X requires one that costs $220+.
Although it's clear that the R7 1700 is considerably cheaper than the Core i7-7820X, we've been wondering just how much faster Intel's solution is considering both chips have 8 cores and 16 threads. Granted, we've already run plenty of Ryzen and Core-X benchmarks here at TechSpot, so we have a pretty good idea of how these CPUs compare.
Initially this article was going to be one of those "for science" type deals where the information would have been pointless for the average consumer, but looked into something most other benchmarks don't. I wanted to see how the Skylake-X-based Core i7-7820X fared against the Broadwell-E 6900K clock-for-clock at 4GHz using the same DDR4-2666 memory.
Figuring you guys would want to see Ryzen 7 in the mix, however, I also tested the 1700 at two configurations: running at 4GHz using DDR4-2666 as well as DDR4-3200 memory. For the fullest picture possible, I then went back and re-tested the 7820X at 4.5GHz with a 3GHz mesh overclock and DDR4-3200 memory.
By this point I was heavily sleep deprived and moved on to write the half-coherent introduction you just read. Looking back, I admit things got away from me a bit here, but the good news is we have 19 graphs and 6 CPU configurations to check out, so the real fun is just about to begin.
Ryzen System Specs
Skylake-X System Specs
Broadwell-E System Specs
Rather than start with productivity benchmarks (see page 3) we're going to skip the appetizers and head straight for the dessert on this one with gaming results, but first let me explain what's going on with all these yellow bars.
The three yellow bars at the bottom represent the Ryzen 7 1700, Core i7-6900K and Core i7-7820X all clocked at 4GHz using DDR4-2666 memory with the same primary timings. The two sets of blue bars compare the R7 1700 and i7-7820X, again at 4GHz but this time with DDR4-3200 memory. Then the golden bars at the top are showing the i7-7820X in all of it's glory at 4.5GHz using DDR4-3200 memory with the mesh interconnect overclocked to 3GHz. Each configuration was tested with the GeForce GTX 1080 Ti.
Looking at the apples to apples results (the yellow bars) for Battlefield 1, we see that the 6900K is a few percent faster than the new 7820X in this title when all things are even, while the 7820X is 13% faster than the 1700.
What's interesting is that once we increase the R7 1700's memory speed it can roughly match the 7820X (we also saw this in our 30 game benchmark when comparing the Ryzen 5 1600 and Core i7-7800X). In fact, the increased memory speed does nothing for the Intel CPU in this game, as the 7820X is now just 2% faster when comparing minimum frame rates. Winding up Intel's 8-core CPU to 4.5GHz didn't help that much with the 7820X being only 3% faster than the humble R7 1700.
Civilization VI is a game where Ryzen has looked quite poor in the past next to Intel's quad-core offerings such as the 7700K and it's a little sluggish even to this day. That said, throw Core-X into the mix and Ryzen looks fine.
Clock-for-clock with the slower 2666 memory, Ryzen was 11% faster than the 7820X though shockingly 14% slower than the 6900K. Increasing the memory frequency to 3200 boosted the 7820X performance by a whopping 17%, though less whopping when compared to the R7 1700's almost 30% performance increase. We're only seeing a 20% boost in memory frequency here so the 2666 speed must have been imposing a serious bottleneck for Ryzen.
Overclocking the 7820X to 4.5GHz helped to extract a further 11% performance but even to it still trailed the 4GHz R7 1700 by a 9% margin. Looking back at our 30 game benchmark, Civilization VI provided the 7800X with one of its worst results against the R5 1600, so let's move on to see how things go in F1 2016.
Performance is much more competitive here. Looking at the DDR4-2666 numbers, the 6900K was actually faster the 7820X when matched clock-for-clock, and the 7820X was only 6% faster than the R7 1700 when looking at the minimum frame rate.
Upping the memory speed to 3200 once again does little to nothing for the 7820X. The gains for Ryzen weren't huge here either but it was enough for the R7 1700 to roughly match Intel's new 8-core CPU. Nonetheless, our maximum overclock on the 7820X did place it firmly in the lead at around 6% faster than the 1700.
Benchmarks: Far Cry Primal, Hitman, Warhammer
Moving on we have Far Cry Primal which might seem like an odd choice given that it isn't a well optimized game for high core count CPUs, but it often delivers interesting results so I thought it was worth a look. Here we see when using the DDR4-2666 memory, the 7820X is 13% slower than the 6900K, but 8% faster than the R7 1700, at least when looking at the average frame rate performance.
The DDR4-3200 results show that faster memory increased the minimum frame rate for the 7820X by 5% while the R7 1700 enjoyed a 10% boost. The Ryzen CPU was now 8% faster when comparing the minimum frame rate and a few frames faster for the average.
After squeezing everything we could out of the 7820X, it saw an 11% leap in performance and overtook the R7 1700.
Hitman is another title where the older Core i7-6900K slays both the Ryzen 7 1700 and Core i7-7820X when matched clock-for-clock. The R7 1700 made out quite poorly with the DDR4-2666 memory this time, being 10% slower than the 7820X and 17% slower than the 6900K when comparing the minimum frame rate.
Increasing the memory speed to 3200 didn't do much for the 7820X and we've seen this several times already. The R7 1700, on the other hand, enjoyed a 13% jump in minimum frame rate and a 15% increase for the average, so it's not far behind the 7820X here.
Turning up the heat with the Core i7-7820X at 4.5GHz helped improve performance by a further 8%, placing it comfortable ahead of the R7 1700.
Finally we have Total War Warhammer and this is a title that has been well optimized to take advantage of Ryzen through to a few handy updates. As you can see even with DDR4-2666 memory Ryzen is able to lay waste to not just the 7820X but also the 6900K in this title. It's also worth noting that the Broadwell-E CPU was again faster than Intel's new 7820X, delivering an 8% greater minimum frame rate.
Upgrading to DDR4-3200 did improve the 7820X's minimum frame rate result by 10%, which is decent, but the R7 1700 saw a massive 16% performance bump here and stayed slightly ahead of the 7820X even when it was pushed to 4.5GHz and 3GHz mesh.
Benchmarks: Applications & Synthetics
Before checking out memory performance and power consumption, I ran a few synthetic and application benchmarks. Cinebench R15 is good for measuring raw CPU performance and memory performance has little impact here. Interestingly, the Core i7-7820X outpaced the 6900K when matched clock-for-clock, albeit by only 4% faster, but this bucks the trend we've seen so far.
When it comes to multi-threaded performance, the Ryzen 7 1700 and 7820X are on par, though the Intel CPU does enjoy a 5% advantage in single-threaded scenarios. Once overclocked to 4.5GHz, the 7820X pulled well ahead and achieved a score of more than 1900pts.
Next up we have compression and decompression performance with 7-Zip. From what I've gathered, Ryzen's SMT feature helps massively with decompression work but isn't utilized for compression. I haven't looked into this properly yet but whatever the case, Ryzen is worlds better at decompression than it is compression, though that's not too concerning seeing as most users do significantly more decompression work anyway.
Clock-for-clock, the 6900K and 7820X are similar in this test while Ryzen was noticeably faster for decompression but significantly slower for compression. Increasing the memory speed helped Ryzen a bit but had little impact on the 7820X. After being overclocked to 4.5GHz, the 7820X took a large step forward and managed to match Ryzen's decompression performance.
The Blender render time is measured in seconds, so lower is better here. Memory frequency also has little to no impact on performance so this didn't help Ryzen close the gap on the Intel CPUs. Ryzen was 8% slower than the 7820X in this test when compared clock-for-clock and 23% slower once the 7820X was overclocked to 4.5GHz -- a pretty solid win for Intel here.
We find a similar story when testing with Corona. The 7820X is a little bit faster than the 6900K and a lot faster than the R7 1700. With both CPUs overclocked to the max, the 7820X was 16% faster, though of course it does cost more than twice as much so this is hardly a win in terms of value.
Benchmarks: Memory Performance
Before we get to the power consumption figures, here we see that Ryzen's dual-channel memory controller has limited bandwidth compared to Broadwell-E and Skylake-X. Whereas the 7820X can push 66GB/s for the read throughput, the R7 1700 was limited to 40GB/s. That gap only gets wider with overclocking, after which the 7820X hummed along to the tune of 81GB/s.
What's interesting to note is that the Core i7-6900K is significantly better than both the R7 1700 and 7820X when it comes to memory latency. Ryzen does improve with higher clocked memory, as does the 7820X, but that's to be expected.
Ryzen is well down on the graph when looking at L1 cache throughput -- it's basically half that of Intel's CPUs.
That said, while down on bandwidth, latency is much the same.
Ryzen's L2 cache performance is excellent, smashing the 6900K clock-for-clock and even beating the 7820X. At 4.5GHz the 7820X does pull ahead but even so Ryzen is strong here.
Although its overall throughput is weaker, the 6900K's L2 cache offers considerably lower latency than the 7820X until the latter is overclocked. Ryzen couldn't quite catch the 6900K here but still offered strong results.
When compared to the 6900K, the R7 1700's L3 cache throughput looks excellent, but not so much when seated next to the 7820X.
Throughput isn't everything of course and here we see that despite its big bandwidth, the 7820X's latency is lousy even when overclocked. In terms of responsiveness, Ryzen has a big advantage here.
Power Consumption & The Verdict
Since memory frequency has little to no impact on overall system power consumption, I only included the 2666 RAM figures here along with the 7820X's 4.5GHz overclock.
Clock-for-clock, the Core i7-6900K was very efficient pushing total system consumption to just 206 watts in Cinebench R15's multi-threaded test. The Ryzen 7 1700 also performed well at 248 watts while the 7820X was a bit hungrier, hitting 268 watts.
Once overclocked to 4.5GHz, the 7820X increased total system consumption by 26% and pulled 36% more power than the R7 1700.
We have some interesting results to discuss. Let's start with the Core i7-6900K and 7820X.
It was shocking to find that when comparing clock-for-clock performance using the same memory speed on both setups, the older 6900K was faster in every single game we tested and significantly so in titles such as Civilization IV. Even if we give the 7820X the advantage of having faster DDR4-3200 RAM, which the 6900K doesn't support, it was rare for the Skylake-X CPU to take the lead.
When overclocked to 4.5GHz with a 3GHz mesh frequency, the 7820X was still only able to match the 6900K in most of the titles tested and realistically we could have squeezed a few hundred MHz more out of the Broadwell-E CPU. Adding insult to injury, the 7820X consumes significantly more power to deliver similar performance of the previous generation part, not to mention that you can expect to require a high-end liquid cooling setup to achieve the 7820X's 4.5GHz overclock without heavy throttling.
When it came to application performance, the 7820X did look much better, though even then it wasn't always superior to the 6900K. For example, we saw similar performance in 7-Zip, while Cinebench R15's numbers weren't drastically different. The 7820X was marginally better in our Blender and Corona tests but not to the degree where you would find yourself getting excited about the results.
The only advantage the 7820X has over the 6900K is the fact that it's around 35% cheaper ($600 versus $1,050). That's obviously a big deal, but if you made me choose between these two CPUs at the same price, I'd probably take the 6900K.
At $290, it's pretty clear that the Ryzen 7 CPU is in a different league when it comes to value and I don't think the most loyal Intel fanboy could argue otherwise. Factoring in the cost of a motherboard ($220+ versus ~$100), the 7820X is around 130% more expensive than the R7 1700, and of course it was never anywhere near that much faster.
When comparing the R7 1700 against the 7820X in terms of maximum overclocked gaming performance, the results were much the same overall. The 7820X enjoyed a win in Hitman while the R7 1700 was noticeably better in Civilization IV and the rest of the games were largely a wash. However, the 7820X was 23% faster in Blender and 16% faster in Corona, so it was hands down faster for these workloads, just not 130% faster, and to achieve that extra performance it consumed 36% more power.
- Ryzen 7 1700 on Amazon, Newegg
- Intel Core i7-7820X on Amazon, Newegg
- Intel Core i7-6900K on Amazon, Newegg
In the end, the Skylake-X architecture is at best a side-step from Broadwell-E and this has afforded Ryzen a great deal of breathing room. If Intel were even the slightest bit aggressive with its pricing then Ryzen would have a serious fight on its hands, but it currently seems crazy to spend $600 on the 7820X when you can get comparable performance for less than half the price.