AMD Zen processors...Are you interested?

[Life-Binder]

And the gaming crowd, although enthusiastic, is a minority.

its far bigger then the extremely niche video editors/content creators

anyway, as a PC gamer theres no reason to buy a Zen for me .. its slower now in games, it will still be slower after tweaks and it will be even slower against Skylake-X & Coffee Lake, both of which will be out well before Zen+ .. hell, Zen+ is slated for 2019 IIRC, even Icelake has a chance of coming out earlier than Zen+

I can only come up with one scenario in my mind where someone would benefit from a 7700k

every single game in existence right now

The 1700 almost makes the entire X99 platform pointless.

only if you're on a tight budget

if you can afford the best then X99 (6900K) has about parity or slight lead vs Zen in professional apps, OCs ~300+ Mhz higher and has significant leads in games


although theres no point buying X99 now anyway, might as well wait for X299




AMD remains the budget vendor, good for them .. but I dont like to compromise if I dont have to and i7 gives me that option

[Deleted]

I honestly dont get the hate for ryzen, its basically people that are unable to use their brains. They are offering you a 8 core 16 thread CPU that is a perfectly capable gaming cpu as well for 330 bucks, that can overclock on sub 100 dollar motherboards.

In time you guys bashing ryzen will understand how good of a product this is for the money. Its basically sandy bridge v2, its an absolute game changer.

Remember guys i havent owned an AMD product in ages, and have never been so excited about a PC upgrade.

I am pretty excited for Ryzen, I currently run a 5820K clocked at 4.3 Ghz with tweaked voltages as well and think the Ryzen chips have more potential, people can only function with the mindshare, the performance is good and I am honestly going with the Jokers results here, he ran those games exactly like people run their games with, it was representative of the performance you would get when running the games.

What people don't realise is, to run HEDT, you need to invest in expensive motherboards and Ryzen is giving you a cheaper option for close the performance if not the same.

[Life-Binder]

Ryzen is about the same as Haswell refresh in gaming with improvements coming.

it loses to 4790K in almost every game, see gamers nexus benchmarks

[Deleted]

its far bigger then the extremely niche video editors/content creators

Ok thats wrong, want to count how many youtubers there are again alone?

Want to take a look at how many digital artists there?

Want to take a look at how many CAD designers there are?

Want to take a look at any form of content creation artists there is anyways?

Gamers who have anything above a dual core is tiny in numbers, hardware vendors make money in the professional sector since you also need to factor in Apple users, shit ton of design studios use Macs dude.

- - - Updated - - -

it loses to 4790K in almost every game, see gamers nexus benchmarks

And pointing to the Jokers results in 1080 P shows different results, Nexus always has different results to everyone else no matter what they do, his Doom results in Vulcan is odd for one.

[BicycleMafioso]

Gamers are enthusiasts, enthusiasts are LOUD. It's a small market, but an important one because they are likely to be the ones giving advice to others.

- - - Updated - - -

it loses to 4790K in almost every game, see gamers nexus benchmarks

Good one.
1800x is ahead of 4790k in BF1 720p, equal in Witcher 3 in 720p (but better minimums), equal to the 7700k in Civ VI 720p (but worse minimums).

[Fascinate]

Its just so ridiculous tho lol. The problem is people didn't know what to expect, but if you have kept up with the development of zen you would have tempered your expectations for gaming and realized the platform you are buying into is an absolute steal vs the competition.

The main point id like to make is the experience you get while gaming. I have seen three outlets (as well as people on forum boards) claiming they are getting a better GAMING experience with ryzen than they had on an i7. WHo in the hell stares at a FPS counter while gaming, i consider myself somewhat of an enthusiast and sure while ill open it once i dont leave it open, i dont care. If the game plays smooth, who cares if you are getting 90 fps or 110?

Ryzen is a no brainer chip vs anything intel is offering right now, in time you will all realize this.

[Artorius]

It was already explained in this thread why different reviewers had different benchmark results, GamerNexus was most likely just bottlenecking their sample horribly because they could never make their RAM work correctly as advertised. When a simple BIOS update can give you 25% extra performance, and using Windows 7 also gives you a double digit performance increase, you know the platform currently has problems, and the RAM problem is just one of them. Joker had the good mobo and his RAM kit worked perfectly, his results are all really consistently decent which is pretty telling. He also has a video on 720p low settings gaming that perfectly show how draw call performance is obviously better on Kaby Lake, which was always expected. Kaby Lake is obviously better than Broadwell-E at it too so what's the surprise? The guy isn't biased in his releases and we have no reason to believe he's lying. The amount of useful information you can grab on The Stilt's thread is also backing up everything we've been saying for a while.

[Thunderball]

No kidding. Not to mention I don't understand how people are so thrilled the next generation chip managed to beat out the flagship of the other guy that has been around and on the market for a long time. I see no real race until I see both sides next generation. That is when both sides are aware and directly competing instead of one guys older platform being edged out by the other guys newer platform they specifically targeted to barely edge out. What a huge shocker.

I'm all for competition, it's always good at least in a sense of lowered prices. But it's this case it's just pure hype.

Its, not it's.

It is, however, better for gaming than Broadwell-E unless you specifically want 3 or more GPUs -- which no one would want.
You are also incorrect. Gamers buy mainstream mid-segmented CPUs most often, while the prosumers are the ones who most commonly buy high-end performance processors.

Ryzen is about the same as Haswell refresh in gaming with improvements coming. How anyone think their performance is objectively bad in gaming is stupefying.
People are very excited about Ryzen and for good reasons.

Incorrent to compare Ryzen with Broadwell-E for gaming. Broadwell-E is miles better and is miles more expensive. Also, we're not talking about mainstream gaming here, Ryzen 7 is too expensive for that.

[Life-Binder]

Jokers results

[Deleted]

I'm all for competition, it's always good at least in a sense of lowered prices. But it's this case it's just pure hype.



Incorrent to compare Ryzen with Broadwell-E for gaming. Broadwell-E is miles better and is miles more expensive. Also, we're not talking about mainstream gaming here, Ryzen 7 is too expensive for that.

'Miles better' You seem to be seeing the numbers differently to me here then on the various articles, the performance difference is small and this is someone who owns Haswell-E.

- - - Updated - - -

Who the actual hell plays at those settings

heres the actual benchmark



This is how people actually configure the games to play.

[Life-Binder]

benchmark is meant to tell you which CPU is objectively faster, not how to play your games

[Thunderball]

It was already explained in this thread why different reviewers had different benchmark results, GamerNexus was most likely just bottlenecking their sample horribly because they could never make their RAM work correctly as advertised. When a simple BIOS update can give you 25% extra performance, and using Windows 7 also gives you a double digit performance increase, you know the platform currently has problems, and the RAM problem is just one of them. Joker had the good mobo and his RAM kit worked perfectly, his results are all really consistently decent which is pretty telling. He also has a video on 720p low settings gaming that perfectly show how draw call performance is obviously better on Kaby Lake, which was always expected. Kaby Lake is obviously better than Broadwell-E at it too so what's the surprise? The guy isn't biased in his releases and we have no reason to believe he's lying. The amount of useful information you can grab on The Stilt's thread is also backing up everything we've been saying for a while.

Watch the explanation video from GamerNexus. The guy explains everything just for the hype train riders.
1) They had RAM running at 2933, which is VERY high for Ryzen, considering other reviews. They also used a certificed kit.
2) They had the lastest BIOS update for their mobo.
3) He also explains why Joker's methodology is incorrect, explaining why his results were so much different from other reviews.

In the end, there is no point of benchmarking with a GPU bottleneck (as AMD suggested reviewers do) as it will equalize results. They were reviewing a CPU, not GPU bottlenecking.

EDIT: Hype train rolling CHOO-CHOO-CHOO.

[Artorius]

Watch the explanation video from GamerNexus. The guy explains everything just for the hype train riders.
1) They had RAM running at 2933, which is VERY high for Ryzen, considering other reviews. They also used a certificed kit.
2) They had the lastest BIOS update for their mobo.
3) He also explains why Joker's methodology is incorrect, explaining why his results were so much different from other reviews.

In the end, there is no point of benchmarking with a GPU bottleneck (as AMD suggested reviewers do) as it will equalize results. They were reviewing a CPU, not GPU bottlenecking.

EDIT: Hype train rolling CHOO-CHOO-CHOO.

1 - The frequency isn't the most important thing here, the timings are. The way they configured their RAM probably locked it to a very loose timing which isn't something you can change manually currently. I already explained what's currently wrong with the controller, you can go back and read the thread again, since you didn't. Even better, go watch Buildzoid's video, he explains it better than me.
2 - And the latest BIOS update for their mobo was still bad. Unfortunate that we had BIOS problems on launch, but what can we do about it? Nothing other than using the mobos that were working half decently.
3 - It isn't, Joker benchmarked more than a single scenario and he wasn't in a single one of them GPU bottlenecking it by running in UHD. Joker also showed that when you put ridiculous settings Kaby Lake will obviously be better due to obvious reasons, it has a ~7% edge in IPC and clocks 800MHz higher. It's also better than Broadwell-E so again what's the surprise?

Ryzen is trading blows with Intel's HEDT line, it wins in some cases against Broadwell-E, loses in others. However both lose to Kaby Lake in single threaded scenarios which is something that is simply obvious and I don't know why you'd expect it to be otherwise. It is also doing this at half the TDP, less than half the cost compared to the 6900K and also gives you a safer upgrade path since AMD stick to their sockets for much longer than Intel. The entire X99 platform is pointless unless you need to go with the 6950X, paying 5x more for 30~35% better performance.

[Deleted]

benchmark is meant to tell you which CPU is objectively faster, not how to play your games

Gamers buy hardware to run their games well, the real benchmark shows little difference, this is a real world scenario in which when said gamer buys the part, they will get those results.

What that test does is show a specific test which most people will not replicate and will not not gauge how the gaming load effects performance in various scences, running higher settings actually affects CPU differently to on low as the geometry algorithm changes, not to mention different post effects, FXAA is not the same as SSAA, SSAA requires more drawcalls especially in physics heavy scenarios.

Low settings is far more serial and is more clock speed dependant in which higher settings take advantage more on multicore or architectural improvements, I would be very curious to which CPU does better with the same clock speed, if you notice the I7 7700K is about 20% better FPS in the low benchmark, try and guess how much faster its clockspeed is to the R7 1700.

If you put those settings on higher, you will notice the FPS difference is a lot less despite the clock speed difference staying the same, the benchmark sadly does not tell you whats actually better in games, what it does say is if anything is clock speed sensitive, the I7 is better.

[Life-Binder]

the 7700K is still faster by the same amount on higher settings and res, you just need more GPU power to see it


settings dont magically make zen faster relative to i7

[BicycleMafioso]

Incorrent to compare Ryzen with Broadwell-E for gaming. Broadwell-E is miles better and is miles more expensive. Also, we're not talking about mainstream gaming here, Ryzen 7 is too expensive for that.

No. It's very correct to compare them, because they are the ones they are going up against.
You are very correct in that Broadwell-E is miles more expensive. Which is exactly why Ryzen being so very close to Broadwell-E, which it is, is so very impressive, considering.

For the prosumers, of course, Ryzen slaughters Broadwell-E. In price to performance, in actual performance, in price and in performance per watt.

- - - Updated - - -

benchmark is meant to tell you which CPU is objectively faster, not how to play your games

This is very true and I agree. However, people need to know how to apply and make use of the data being presented.
Low 720p in a game without FPS-cap is how you benchmark a CPU for gaming - to move the focus-point (and people need to stop using the word bottle-neck incorrectly) to the CPU, this is what is needed

But running actual real-world scenarios is what tells you how they actually will perform. And this is why Ryzen not only doesn't suck, it's close to intel. Which is damn impressive, given their situation and what they've had to work with.

[Deleted]

the 7700K is still faster by the same amount on higher settings and res, you just need more GPU power to see it


settings dont magically make zen faster relative to i7

Actually you don't, the low settings had a 20% difference in his test, but the 1080 P results had 10% or less, in fact I would say 5% or less difference, this is consistent with how different graphic settings affect CPU load, like my example of FXAA to SSAA, or even draw distance and FoV.

The fact you don't understand this makes me wonder if you understand PC gaming at all, light loads which only requires basic numbers to be crunched fast as possible has clock speeds as the governing factor, complex numbers and algorithm requires IPC to deal with that as complex effects require more drawcalls.

So yes, settings do change things entirely, SSAA will hit the CPU more but the extra pixels that is used on the geometry requires draw calls, however shadows, lighting, draw distance and physics will require the CPU more and they don't just up the resolution of said effect, they also become more complex in which they require more draw calls again.

[Artorius]

I would be very curious to which CPU does better with the same clock speed

Single threaded with SMT/HT turned off at 3.5GHz for all the uarchs benchmarked:




Multi-threaded benchmarks with 4 cores and no SMT/HT, still 3.5GHz:



Now with SMT/HT turned on:



SKU vs SKU comparison:



SMT/HT yield:



Pretty much all we can learn about it at the moment in a very compressed way, you can check all the benchmarks here to see where it's better and where it's worse against the other uarchs.

[Life-Binder]

the 1080 P results had 10% or less,

1080p Ultra actually starts having GPU bottlenecks fyi, even on Pascal

especially unoptimized mess like WD2 & Deus Ex MD


so back to:

the 7700K is still faster by the same amount on higher settings and res, you just need more GPU power to see it


settings dont magically make zen faster relative to i7

[Dorfadin]

its a nice cpu for sure, big step up in performance compared to amd's other stuff, however the temperatures I've seen in testing, that tinytomlogan from OC3D was benching the 1800x at 82c !!! with a h100i cooler, omg that's way, way too high, cpu would probably die in a years time constantly running, so I will happily stick to my 4790k, same performance in games and only 35c temps under load using a h100i cooler