thelastword
Banned
Preliminary Reviews/ Preview and Leaked Benchmarks
https://wccftech.com/amd-ryzen-9-3900x-and-ryzen-7-3700x-cpu-review-leak-gaming-benchmarks/
Reviews
The AMD Ryzen 7 3700X is an 8 core and 16 thread part featuring the 7nm Zen 2 core architecture. The chip features a base clock of 3.6 GHz and a boost clock of 4.4 GHz. There’s 36 MB of L3 cache, 40 PCIe Gen 4 lanes (CPU + PCH) and a TDP of 65W (derived from the base frequency).
Might as well wait for next genTomorrow is when I will decide I will keep my I4790k for a while longer, or upgrade. This CPU is 5 years old, but doesn't feel like it. Still, 3x the cores at a reasonably low price sounds too good. Still, I would need to get a new Mobo and ram as well, and maybe a new power supply just to be safe, since this one is kind of old.
Genuine question since I don't understand: why are those benchmarks at 720p?Early review can be seen here. 12 Core 3900X trades blows with 30 month old 7700K (Quad Core) in gaming workloads.
Good to see AMD making progress with faster CPUs every year but it appears Intel, with it's recent struggles, still has the gaming crown.
Genuine question since I don't understand: why are those benchmarks at 720p?
At the risk of a lot of downvotes, I figured I would try to explain why testing CPUs at low resolutions is important. I am seeing a lot of misinformation and rationalizing for why we shouldn't test at low resolutions like gamers playing at high resolutions and all, but honestly, when you test CPUs, you actually want to test the CPU. Including benchmarks at high resolutions can be nice to give people who game at higher resolutions an idea how their system performs, but all in all, to actually TEST the CPU, you want to test at low resolutions.
Benchmarking tests are a lot like scientific trials. You want to link an independent variable, to a dependent variable. In this case you want to test how your CPU runs at gaming. What you don't want interfering with your trials is other variables. You want to CONTROL the environment, and ensure that the only factor at work making a difference is the thing you are testing. EVERYTHING ELSE should be controlled for, as it does not interfere with the result.
When you game at low resolutions, you are doing just that. When you test at high resolutions, you are introducing a third variable into the mix: the graphics card. When you test at high resolutions, it's possible that the GPU is taking on more stress from the system, and this could be the primary variable affecting results, NOT the CPU. This gives the viewer the impression that the CPUs are performing more equally than would in a more controlled setting.
To minimize the stress of the GPU, you want to test at max settings (since settings CAN impact CPU performance), and you want to test at as low of a resolution as possible. I'm seeing joking comments asking, what are tests only valid if you test at 640x480 or something? Well, let me put it this way. If you're running a GTX 1080 at 640x480, you're not going to be GPU bound in most games. You would be testing the CPU, and this would lead to more valid results, since that's what we are testing.
This isn't to say that high res benchmarks can't be good. It's good to test at higher resolutions to give people some baseline expectations to how it will perform in a real world environment, but testing at low resolutions is more important, because it tells you how the CPU will perform under stress, and this could give an indication of the long term performance.
Because that's the way we play .......Genuine question since I don't understand: why are those benchmarks at 720p?
Hmmm....................
Lol yeah we should test them the way it's used to play, that's for sure.Because that's the way we play .......
Hmmm....................
Lol yeah we should test them the way it's used to play, that's for sure.
A more practical test would be to test games that punish CPUs at 1080p and upDepending on resolution that's a mostly GPU bound scenario. For example, if you're gaming at 1440p or 4K, the slowest 8 Core first gen Ryzen will be within single digits of the fastest 12 Core Zen 2 Ryzen. Not a good way to test the CPU...
A more practical test would be to test games that punish CPUs at 1080p and up
What is the point of buying a cpu for "top gaming performance" if it will provide the same results as the cheaper alternative for your target resolution.
Sure you could argue is more future proof but the way forward for games optimization is multithreading and as such 8 cores and up will be more beneficial than high clocks that show an advantage at 720p
It was more targeted at people that say they'll get Intel for top gaming performance but if that lead only shows at 720p its a mute point.Simple. Because the test was conducted with a single 1080ti. That's not the maximum amount of graphical horsepower available so unless you're going to find two 2080tis, an easier way to reveal the true upper limits of the cpu is to bump it down to 720.
And it's a review. They're supposed to be scientific. Seems a little biased to come to a reviews thread and then lament the fact that the testing method wasn't more practical and didn't focus on 1080p AND up (as if esports at 1080p 240+fps doesn't matter).
BTW, I'm buying a 3700x in all likelihood. I think it's a waste of a comment/disengenous to suggest we don't look at 720p because that's 'not the way forward' or some other limiting bs. Not everyone plays AAA open world games exclusively smh
and no security risk in leaving HT enabled either.Either way, the 3700X looks like the big winner here. Solidly competes with the 9900K and over $100 less.
I can't even imagine what devs will cook with 7 zen 2 cores (assuming 1 reserved) and 36MB cache.These are going to be a bigger game changer in the consoles than I think people realize.
Will bode well for PC ports as well.
I don't expect the console counterparts to have the same amount of cache at all. It'll be some form of 'lite' CPU. But even so, yeah I agree. It's gonna be great.I can't even imagine what devs will cook with 7 zen 2 cores (assuming 1 reserved) and 36MB cache.
doesn't need to be, in fall 2020 these cpus will be more than 1 year oldI don't expect the console counterparts to have the same amount of cache at all. It'll be some form of 'lite' CPU. But even so, yeah I agree. It's gonna be great.
I think they might, 8 core Zen 2 chiplet with 36MB L3 is just 75mm2 after allI don't expect the console counterparts to have the same amount of cache at all. It'll be some form of 'lite' CPU. But even so, yeah I agree. It's gonna be great.
Well the Threadripper 2950 is in last on that FarCry 5 figure... so it's one hell of an upgrade. Also, the i9 7980XE (18c/36t) is behind the 7700 also.Losing out to the 4 year old Skylake 7700k in many cases is concerning to me.
Can f$%¨* wait.AMD Ryzen 9 3900X 12 Core and Ryzen 7 3700X 8 Core CPU Review Leaks Out - Gaming Performance On Par With Intel's 9th Gen, Ryzen 9 3900X Consumes Less Power Than i9-9900K
Another AMD Ryzen 3000 series review has leaked online and this time we are looking at the 12 core Ryzen 9 3900X & the 8 core Ryzen 7 3700X.wccftech.com
Wasnt it 32MB? This CPU will be beast in consoles even at 3.2GHz
There's absolutely ZERO security risk of leaving HT enabled. Literally no one give a fuck about you, about your PC and never will, just like for the last 10 years of so. All of this "Security" nonsense in regards to Intel's CPUs is pure bullshit and normal / regular people should just ignore it completely.and no security risk in leaving HT enabled either.
Where are these CPUs? No review and nobody is selling them.
????????
still under embargo/nda. should be another couple hours i think.Where are these CPUs? No review and nobody is selling them.
????????
Yeah CPU reviews WRT gaming have kind of been a joke for a long time. They turn the resolution way down to emphasize the CPU differences.A more practical test would be to test games that punish CPUs at 1080p and up
What is the point of buying a cpu for "top gaming performance" if it will provide the same results as the cheaper alternative for your target resolution.
Sure you could argue is more future proof but the way forward for games optimization is multithreading and as such 8 cores and up will be more beneficial than high clocks that show an advantage at 720p
4790k will be good until next gen.Watched a couple of reviews and it seems that the 9900k slightly outperforms the 3900X in terms of gaming, which means it's going to be about as useful as my 4790k, at least for now. Microcenter has the 9900k for slightly cheaper than the 3900x, so it makes little sense to drop the money for either one now.
Either 4790k is an absolute beast, or CPU's have just been at a complete standstill for awhile.
That's just it. Measuring at 4K becomes a GPU test.I really wish these reviews would talk about 4k performance. I know that at 4k, it becomes mainly gpu bound, but I'd still like to know.
SameI really wish these reviews would talk about 4k performance. I know that at 4k, it becomes mainly gpu bound, but I'd still like to know.
Linus review :O
In gaming performance the non-overclocked 3700X lags around 3-5% in the 9900K, but is $150 cheaper. There is no "eh?" about it. Thats $150 savings can go towards other parts. $150 is the difference between 2060 and 2070 Super. $150 could be put towards getting a 2080 Super or Ti. $150 is the difference between being stuck with a HDD and a M2 drive.Same
I want the i9 9900k to be beaten and good this is... well eh?