Ryzen Awesome True Look

[Prysin]

1 minute ago, Misanthrope said:

Actually I must call this a false dichotomy.

 

A person that is considering this 2 systems could be considering very different use cases:

 

1) Person running 7700k + 1080 is probably a gamer

2) Person considering 1700 + 1080 is probably a budget workstation user.

 

For person 1 there is no reason to consider the option 2). For person 2 he is weighing his option vs the 6800k and above not the consumer chips. Just because he puts a "gaming" card in there doesn't means he wants it for gaming (or only for gaming) since it's still a great, cheap way to get CUDA support at a fraction of the cost of going for a higher end workstation build with a 6950k and a Quadro card.

 

So can someone game at budget 2) Surely. But that someone that's just gaming at that budget would be much better off going for a 7700k + 1080ti rig instead.

i would call that false presumption.

I got a idiot friend who bought a 5960X + Dual Titan X (maxwell) SLI... he spends most of his day playing Nintendo WII and watching anime. When he does use his PC its for old games like Baldurs Gate and shit like that....

 

Many people DO buy overpriced shit for the e-peen factor.

Other people do like me and buy a i7 over an i5, because we KNOW that in a few years, that i5 will progressively suck, while the i7 will retain its performance more gracefully. Why does the i7 do so? MOAR THREADS. Simple as that. IPC and frequency aside, Hyperthreading/more threads always beats out less threads in the long run.

[Misanthrope]

4 minutes ago, Prysin said:

uhm... I cannot speak too much for DX12, but Vulkan is NOT that hard to optimize of VERY many threads. Iv'e invested in the official vulcan coding book and looked heavily into learning to code for Vulcan (if only for lulz). And from what i can see so far, while being unable to APPLY the code in an appropriate or representable way, i can tell you it is not rocket science.

 

Any dev worth a damn shit could pick up the official vulcan text book from Amazon for 45$ and he would know within a few hours how to do it.

This gets off topic but there's a lot of BS passed down by the suits that get flatout money hand outs to include shit like Gameworks so honestly, for business and not technical reasons, I don't really see Vulkan taking off in the near future.

[Misanthrope]

2 minutes ago, Prysin said:

i would call that false presumption.

I got a idiot friend who bought a 5960X + Dual Titan X (maxwell) SLI... he spends most of his day playing Nintendo WII and watching anime. When he does use his PC its for old games like Baldurs Gate and shit like that....

 

Many people DO buy overpriced shit for the e-peen factor.

Other people do like me and buy a i7 over an i5, because we KNOW that in a few years, that i5 will progressively suck, while the i7 will retain its performance more gracefully. Why does the i7 do so? MOAR THREADS. Simple as that. IPC and frequency aside, Hyperthreading/more threads always beats out less threads in the long run.

Considering idiots into the equation is not really conductive of anything worthwhile: The fact that you can sell pet rocks to people shouldn't be considered: I am more than certain that if AMD had a gigantic marketing budget they could turn around their company with nothing but the old Piledriver CPUs still just by doing a bunch of invasive and shady shit along with a very massive marketing campaign.

[Sauron]

Just now, alexcheetah said:

 

Did you seriously just post a video comparing the pentium to an i5 when we were talking about athlons? Do you realize the i5 was 3 times more expensive?

 

Please read the posts you quote, it will save everyone's time and sanity.

[Prysin]

1 minute ago, Misanthrope said:

This gets off topic but there's a lot of BS passed down by the suits that get flatout money hand outs to include shit like Gameworks so honestly, for business and not technical reasons, I don't really see Vulkan taking off in the near future.

Vulkan will be taking off. Because of mobile and probably also next gen consoles?

i would NOT be overly surprised to see Sony use Vulcan for next gen consoles. Simply because its free, its open and most importantly, the install base is BIGGER.

 

The number of clients who can buy Vulcan games, VS DX12 games, are simply WAY bigger. Because Vulcan works across all platforms (except Mac, for fruity reasons), across most versions of Windows and on hardware dating back to Fermi/Tahiti.

[Vode]

4 minutes ago, Sauron said:

Oh I read your posts, and your main point seems to be based on the idea that windows bugs will get fixed and games will be optimized. Once again, this is nothing but speculation. If nothing changes, which is an assumption that MUST be made when running benchmarks, then those low res benchmarks will become more and more relevant at resolutions people actually run their games at.

 

Sure, it's not a guarantee of anything, but as @Misanthrope pointed out by your reasoning we might as well stop running benchmarks alltogether because they are not absolutely reliable and indicative of how things will be in the future.

 

As much as I'm optimistic about ryzen in the next few years, you have to consider that EVERY TIME that this sort of argument (it will get better in the future!!!) has come up it has been proven wrong by the market, without exception as far as I remember. This happened with the pentium 4, it happened with bulldozer, and it MIGHT be happening now, even though ryzen's shortcomings are not nearly as bad.

Not entirely. Low resoltuion game benchmark are the same kind of speculation about future performance or can you backup your argument with benchmarks? Adored provided some to show the opposite.

 

The issue I have was that the omitting of the core count and firmware/software side of things while focusing on low resoltion benches as backup for an argument.

 

You dimiss one side of the argument but overemphasize on the other, dismissing one as invalid while saying one is scientific and accurate when it isn't.

 

Man I'm not saying benches are worthless, if you play CSGO at 240Hz get a 7700K but this is not indicative of real world gaming performance but let's not get sidetracked because this isn't my argument at all.

 

PS: Just three cases where software made it better. 8350 was beat by 2500k in 2012, now it's the opposite, 780ti > R9 290 at lanch, now quite the opposite. GTX 1060 > RX480 at launch, now they are equal.

[Misanthrope]

Just now, huilun02 said:

Your argument is even more of a false dichotomy... blanket classifying each setup for distinctly different use case.

And then detracting from the fact that the bottleneck is on the GPU for most people, by saying a gamer would have a 1080ti.

An R7 1700 would be great over the 7700K for gaming and general use simply because people aren't running test systems with flagship GPU and fresh Windows install.

I edited my post a bit after re-reading: Didn't mean to assert there's 2 cases but that there could be 2 more cases.

 

I have not seen a single reviewer not using a fresh windows install and I see no point in using a GPU to test CPU performance: the CPU should be isolated with 1080p tests with a decent GPU that can handle the resolution. What you're suggesting is cherry picking to make AMD look closer in overall IPC performance than it really is, that's not productive for people not looking to spend money on unnecessary extra cores for their use.

 

You could just suggest waiting for the R5 Ryzen chips since they might clock and overclock higher thus addressing the IPC deficit complains here and at a lower price to boot.

[Misanthrope]

3 minutes ago, Prysin said:

Vulkan will be taking off. Because of mobile and probably also next gen consoles?

i would NOT be overly surprised to see Sony use Vulcan for next gen consoles. Simply because its free, its open and most importantly, the install base is BIGGER.

 

The number of clients who can buy Vulcan games, VS DX12 games, are simply WAY bigger. Because Vulcan works across all platforms (except Mac, for fruity reasons), across most versions of Windows and on hardware dating back to Fermi/Tahiti.

I kinda doubt that. I mean yes but the use case of performance gains on mobile devices is very different it's not about visual fidelity but just getting playable framerate on crap hardware.

 

But I don't think that Sony would be launching a new PS4 they'll do another "Plus" or "Pro" iteration and continue with incremental updates, essentially becoming a walled garden PC.

[Sauron]

1 minute ago, Vode said:

Man I'm not saying benches are worthless

 

Your own words:

Quote

720p benches are worthless as an indicator of future performance

from this it certainly didn't look like you were specifically referring to situations where low res benchmarks are used as the ONLY evidence or anything like that, and you didn't clarify in your further posts.

 

Now, aside from the 8350 (which got better WAY too late for it to make any difference) every other example you made (on gpus by the way, which are a different matter entirely) was already a good product for its price when it was launched. And while I think that overall ryzen is a good product, it's not very competitive in the gaming market.

[RagnarokDel]

54 minutes ago, Misanthrope said:

 "In the near future games will benefit more from many cores!" and so on and so on.

I still have my core2duo, it clocks to 4.2 so I'll put it against my 4670k at that I'll downclock to 2.5 ghz, guess which CPU is going to win benchmarks.

 

PS: Cant benchmark Ubisoft games cause they literally dont work with 2 cores anymore but every other game!

[Misanthrope]

1 minute ago, RagnarokDel said:

I still have my core2duo, it clocks to 4.2 so I'll put it against my 4670k at that I'll downclock to 2.5 ghz, guess which CPU is going to win benchmarks.

 

PS: Cant benchmark Ubisoft games cause they literally dont work with 2 cores anymore but every other game!

Great: We'll take a look in 2022, not today.

[Misanthrope]

1 minute ago, huilun02 said:

I am not suggesting cherry picking of any sort. IPC is set in stone and we can leave it at that.

What I am saying is that in actual use case, people have shit running in the background,

You're asserting this based on nothing: Which reviewers are you ready to accuse of improper benchmarks? How are you gonna confirm that performance discrepancies are based on that and not on other platform issues (whenever or not they get resolved)?

 

What you're doing is basically accusing anybody not posting results that serve your case as "Fake news". No, tell us how you know that shit is running in the background.

[i_build_nanosuits]

i'm tired of reading those ''Yeah but who game at 720p with a GTX 1080 blah blah blah'' that's beside the point...low-resolution benchmarking is ESSENTIAL to remove the GPU as a potential limitation...you DO want to see the CPU performance in GAMES then let the CPU RIP through those frames and look how fast it can go...end of the story...and YOU DO want the best performing CPU for GAMING because in the long run you'll be upgrading your graphics card 2 or 3 maybe 4 times before you change platform and AT SOME POINT in the future the CPU will become your limiting factor to performance EVEN THOUGH it can push 180FPS right now in a couple years time once the games gets advanced enough IT WILL dip in the lower FPS margin and it WILL limit your brand new graphics cards.

 

SO, take your head out of the sand for one second and think about it...

[Misanthrope]

Just now, huilun02 said:

I am not accusing anyone of posting unfavorable results. I am just saying that people are not reviewers and that review comparisons should not be taken as actual in-use results. The way reviews are done is like comparing an F1 car against a road car. No shit sherlock an F1 car is going to have a higher top speed. But are car buyers competitive formula one drivers? No.

Since we established the purpose of isolated IPC and CPU tests I am gonna call this asked and answer. It's clear you disagree and want to push AMD's narrative of "It's good enough for 1440p! Therefore pay us more than a chip that does the same from intel for a gamer".

[79wjd]

7 minutes ago, i_build_nanosuits said:

i'm tired of reading those ''Yeah but who game at 720p with a GTX 1080 blah blah blah'' that's beside the point...low-resolution benchmarking is ESSENTIAL to remove the GPU as a potential limitation...you DO want to see the CPU performance in GAMES then let the CPU RIP through those frames and look how fast it can go...end of the story...and YOU DO want the best performing CPU for GAMING because in the long run you'll be upgrading your graphics card 2 or 3 maybe 4 times before you change platform and AT SOME POINT in the future the CPU will become your limiting factor to performance EVEN THOUGH it can push 180FPS right now in a couple years time once the games gets advanced enough IT WILL dip in the lower FPS margin and it WILL limit your brand new graphics cards.

 

SO, take your head out of the sand for one second and think about it...

That's not entirely fair. If multithreaded performance improves then a 1700 is a better choice in the long run. However, devs are just as likely to optimize for single threaded performance. Choosing one over the other is a personal gamble, but right now a 7700k is better, and that's the only thing that can be said with certainty.

 

There is a good chance I would take the gamble and go with the 1700, but it's still a gamble.

[i_build_nanosuits]

1 minute ago, djdwosk97 said:

That's not entirely fair. If multithreaded performance improves then a 1700 is a better choice in the long run. However, devs are just as likely to optimize for single threaded performance. Choosing one over the other is a personal gamble, but right now a 7700k is better, and that's the only thing that can be said with certainty.

Watchdogs 2 and battlefield 1 both use ALL 16 threads on the Ryzen chip you can see them loads spread out completely and yet it's still slower than an i5-4690K in those games.

[i_build_nanosuits]

HAHAHA that guy is a moron...the old ass FX-8350 is faster than an i5-2500K in modern titles...hahahahaha that's very very funny...nice try mate.

[Prysin]

32 minutes ago, Misanthrope said:

I kinda doubt that. I mean yes but the use case of performance gains on mobile devices is very different it's not about visual fidelity but just getting playable framerate on crap hardware.

 

But I don't think that Sony would be launching a new PS4 they'll do another "Plus" or "Pro" iteration and continue with incremental updates, essentially becoming a walled garden PC.

no, PS5 will be a thing. They are waiting for ZEN + VEGA custom SOCs. Which, since they are console SoCs, need to be at server/laptop level power efficiency. Rumors say AMD APUs (desktop/mobile) wont arrive until H2/H1 2018... which makes sense. Desktop APUs should be 1-2 years ahead of a next gen console, as they need to make the custom chip (3 month tape-out mass production + 2-3 years headstart for devs to make games)....

 

Sonys Playstation division is actually one of their only profitable branches. Thus anything to maximize that profit will be done.

[Prysin]

3 minutes ago, i_build_nanosuits said:

HAHAHA that guy is a moron...the old ass FX-8350 is faster than an i5-2500K in modern titles...hahahahaha that's very very funny...nice try mate.

you need to post more now that Patrick is gone. You're definetively the second best source of lulz on this forum.

[plus]

Stockfish and Komodo look promising on Ryzen:

http://www.hardware.fr/articles/956-14/ia-echecs-stockfish-komodo.html

Other sites report pretty useful Fritz 9 performance.

 

Good news for chess

 

[JoeCoke]

1 hour ago, Misanthrope said:

That's what I said from the beginning: I didn't feel like getting to the point of the video since it started out with such stupidity.

Or what you thought was stupidity because you didn't get the point of the video.

[i_build_nanosuits]

3 minutes ago, Prysin said:

you need to post more now that Patrick is gone. You're definetively the second best source of lulz on this forum.

EVERYBODY knows that the intel core i5-2500K flatout DESTROY the AMD FX-83XX in gaming...new and old titles.

[Prysin]

3 minutes ago, i_build_nanosuits said:

EVERYBODY knows that the intel core i5-2500K flatout DESTROY the AMD FX-83XX in gaming...new and old titles.

thank you. Please keep going. My day has been dull as fuck so far. You're certainly bringing a good laugh.

[Jahramika]

1 hour ago, Misanthrope said:

I agree but that was kind of my point: we haven't seen true potential for DX12 and Vulkan since it's just not easy to code for it. Whenever or not it's 2 to 3 years like you suggest or longer than that it remains to be seen at this point henceforth making purchase suggestions based on future performance kind of unreasonable. We can only hope that you're right but nobody knows if DX12 or Vulkan will eventually deliver on their promises of greater CPU utilization and optimization.

If you watched the video he showed you that in BF.

[i_build_nanosuits]

10 minutes ago, Prysin said:

thank you. Please keep going. My day has been dull as fuck so far. You're certainly bringing a good laugh.

Yo MAN...check this out ok...EVEN leaving the whole amd FX crap out...

this is his main slide for the ''NOW PERFORMANCE IN GAMES'' whatever ok...check this out

Since when the multicore intel scrap chip beat the 7700K in games by 9%?!

Since when the 4770K is 28% slower than a 7700K?!

since when the 7600K is 8% nfaster than the 4770K (ESPECIALLY IF THE MULTICORE SCRAP ARE FASTER THAN 7700K...BS RIGHT)

since when the 2500K is 42% SLOWER than the 7600K (there is less than 25% IPC difference between them)

 

This guy is BULLSHIT...he look at BULLSHIT reviews from BULLSHIT websites and then post a BULLSHIT video about it...WAKE UP SON.