Re: AMD Ryzen 7 1700X (14nm Zen)
Ideally we'd make the reviewers work harder by insisting on various resolutions and settings even for CPU reviews.
Some of the higher quality settings do use CPU resources so running at 640x480 or 800x600 in low quality settings doesn't necessarily always show how well a CPU can cope with some theoretical monster drawcall load in the future.
Another thing reviewers do not cover is having any kind of streaming going on (ShadowPlay, ReLive etc.).
Also, multiplayer is very tricky to benchmark. This goes back to BF4 and Mantle, where benchmarks noted only small differences but yet people playing multiplayer noticed huge differences with smoothness.
Someone over on the AT forums posted this:
https://i.imgur.com/sZCG0N5.png
Re: AMD Ryzen 7 1700X (14nm Zen)
I don't see benchmarks as an indication of what to expect at X or Y resolution, to me they are an indication of how much better or worse X is when compared to Y, there are to many variables involved with PCs to be able to compare one system with another.
Re: AMD Ryzen 7 1700X (14nm Zen)
ordered mine today .. asrock taichi and a 1700 .. with 3333 ram .. plus a few other goodies ..now to wait for vega
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
anselhelm
Thank you for your sensible QHD (1440P) gaming benchmarks in this review :)
A lot of sites have been annoying benchmarking at ridiculously-low resolutions like 640x480 or 800x600 in order to show a massive Intel bias. I wish I were joking here, but sadly I am not: this is what sites have really been doing.
Other sites have pointed out that AMD is technically "losing" still at 1080P, but that anything above that shows no discernible difference between AMD and Intel processors (even when paired with the highest-end gaming cards available).
Any chance you could do a review on the cheaper R7 1700 now too?
Hi there,
As soon as we have the R7 1700, which should be in the coming week, it will be on page.
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
Tarinder
Hi there,
As soon as we have the R7 1700, which should be in the coming week, it will be on page.
According to XS forums testing,they saw the R7 1700 at similar voltages to the R7 1700X/1800X consume less power??
Any chance you could see if that is true - the hypothesis is that the R7 1700X/1800X are leakier parts made for more extreme cooling and the R7 1700 should do better under normal cooling.
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
Tarinder
Hi there,
As soon as we have the R7 1700, which should be in the coming week, it will be on page.
Do you know I kinda miss the Tables of Doom ?
(But then I am an Accountant !)
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
Firejack
We learn very little about CPU performance by benching in GPU-limited scenarios. I agree that 640x480 is irrelevant as next to nobody uses those resolutions anymore. However 1080p should always be included in any gaming bench. 1080p accounts for 43.23% of Steam users while 1440p is only 1.81% (
http://store.steampowered.com/hwsurvey).
We need results for 1080p and 1440p because it will show CPU performance clearly when you compare the two. By only including 1440p in this review we don't know what to expect when we upgrade our graphics cards. The GTX 1080 Ti is out now. If we swapped one of those with the GTX 1080 in this test suite we would see a clearer difference in CPU performance as the GPU bottleneck is lifted.
And thats what we need to know in this Ryzen 7 1700X review - CPU performance in real world gaming scenarios.
A 1080 at 1080p is not a real world gaming scenario. If we benchmark at a sensible resolution for the graphics card we learn whether the CPU is a significant bottleneck, which is what is important - every graphics card has been GPU limited at the resolution du jour, and this is not likely to change within the next decade. A 1080ti at QHD is not realistic, because only an idiot would spend almost £700 on a GPU to run a QHD in this day and age - you'd expect it to perform at 4K, so if a 1080ti is used then the tests should be run at 4K.
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
Xlucine
A 1080 at 1080p is not a real world gaming scenario. If we benchmark at a sensible resolution for the graphics card we learn whether the CPU is a significant bottleneck, which is what is important - every graphics card has been GPU limited at the resolution du jour, and this is not likely to change within the next decade. A 1080ti at QHD is not realistic, because only an idiot would spend almost £700 on a GPU to run a QHD in this day and age - you'd expect it to perform at 4K, so if a 1080ti is used then the tests should be run at 4K.
I don't fully agree - people do run GTX1080 cards at 1080P,and from my experience of using a GTX1080,some games like Deus Ex:Mankind Divided I have run can't hit a constant 60FPS at qHD(!).
Its still playable,but I do have framerate dips,and I can understand why people might use the card at 1080P instead which I personally would not do either,but its a data point which is still useful.
Planetside 2 is even more ridiculous - I mean I thought it was just a CPU hog,but some of the shadow settings would hit performance badly on the same card at qHD so I reduced them to the lowest.
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
CAT-THE-FIFTH
Any chance you could see if that is true - the hypothesis is that the R7 1700X/1800X are leakier parts made for more extreme cooling and the R7 1700 should do better under normal cooling.
Leakier parts are only really good under LN2 conditions, remember TWKR? http://www.legitreviews.com/amd-phen...processor_1009
Perhaps the non X 1700 just has a lower default voltage?
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
DanceswithUnix
The tests over at XS forums,had the 1700 and 1700X/1800X at the same voltages and the 1700 consumed less power.
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
The snag, then, is that Intel still has a commanding lead in IPC, which tends to be a key contributor to in-game performance and some everyday workloads. AMD has closed the gap compared to the woeful FX-series, but single-thread proficiency isn't quite what we had hoped. Putting the PiFast result into perspective, the top two Ryzen chips are still playing catch-up with the dated Intel Core i5-2500K, which scored 20.5 way back in 2011.
That statement strikes me as ... too bombastic.
a) All the chips in question have faster clocks (4.2-4.5GHz all-core boost vs. 3.5GHz (3.6 if XFR kicks in) all-core and 3.8GHz 2-core boost for the 1700X). That alone should account for 10% (2-core load vs. i3-7350k or i5-7600k) - 20% (pretty much everything else at 4 or fewer cores at full load) performance deficits in up-to-4-threaded workloads.
b) We still haven't figured out the balance between lack of optimization for Ryzen/very specific optimization for Core vs. actual performance differences in specific tasks.
c) PiFast is a single application, and as such it's a stretch to make it the be-all, end-all benchmark for IPC.
Am I saying OMG RYZEN IS TEH BEST U GUISE!!!!1!!? No. But it stands to reason that performance outliers might be corrected by software, driver or microcode updates.
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
Valantar
Quote:
The snag, then, is that Intel still has a commanding lead in IPC, which tends to be a key contributor to in-game performance and some everyday workloads. AMD has closed the gap compared to the woeful FX-series, but single-thread proficiency isn't quite what we had hoped. Putting the PiFast result into perspective, the top two Ryzen chips are still playing catch-up with the dated Intel Core i5-2500K, which scored 20.5 way back in 2011.
That statement strikes me as ... too bombastic.
a) All the chips in question have faster clocks (4.2-4.5GHz all-core boost vs. 3.5GHz (3.6 if XFR kicks in) all-core and 3.8GHz 2-core boost for the 1700X). That alone should account for 10% (2-core load vs. i3-7350k or i5-7600k) - 20% (pretty much everything else at 4 or fewer cores at full load) performance deficits in up-to-4-threaded workloads.
b) We still haven't figured out the balance between lack of optimization for Ryzen/very specific optimization for Core vs. actual performance differences in specific tasks.
c) PiFast is a single application, and as such it's a stretch to make it the be-all, end-all benchmark for IPC.
Am I saying OMG RYZEN IS TEH BEST U GUISE!!!!1!!? No. But it stands to reason that performance outliers might be corrected by software, driver or microcode updates.
I agree. I think we may even see big improvements in certain benchmarks over the next few months as the platform matures.
For me personally though the bottom line is value for money and this is where Ryzen is the big winner (for the moment).
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
CAT-THE-FIFTH
Quote:
Originally Posted by
Tarinder
Hi there,
As soon as we have the R7 1700, which should be in the coming week, it will be on page.
According to XS forums testing,they saw the R7 1700 at similar voltages to the R7 1700X/1800X consume less power??
Any chance you could see if that is true - the hypothesis is that the R7 1700X/1800X are leakier parts made for more extreme cooling and the R7 1700 should do better under normal cooling.
"Leakier parts"... what does this even mean in terms of a CPU?
Re: AMD Ryzen 7 1700X (14nm Zen)
Quote:
Originally Posted by
Stu C
"Leakier parts"... what does this even mean in terms of a CPU?
A leakier part uses more power but is able to tolerate higher voltages. Ideal for suicide runs under LN2 at crazy voltages.
A low leakage part usually can't reach the same max clock but uses less power. Ideal for modest overclocks or undervolting.
TPU's GPU-Z has a read ASIC quality feature for GPUs and what it says mostly applies to CPUs too:
https://i.imgur.com/pEUCU9d.png
Re: AMD Ryzen 7 1700X (14nm Zen)
Am I missing something or does even the high-end Zen only have 16 PCIe Lanes? Seems a bit small for a new architecture/chipset
Re: AMD Ryzen 7 1700X (14nm Zen)
Technically 20, but for anything short of trying to make a supercomputer there's no real performance difference for consumers. SLI/crossfire will happily drop frames whether or not there's a full x16 lane to each card, and the 7700k only has 16
