AMD's Ryzen is here!


#81

void / null …


#82

Let’s not get down this road, please… :stuck_out_tongue:


#83

You mean this tower, right?


#84

I don’t think this is a productive reply …
To put it differently: AMD cores are just not as strong as Intel cores. I think their strategy is not only putting more cores in but also the best ones they have, they’re just not on par. So Intel has an edge in single core tasks (which there are alot of).


#85

AMD disagrees with you. Wait, are we talking about AMD Ryzen or some other Ryzen?

Consider the example of the AMD Ryzen™ 7 1700 processor. It has a base clock of 3.0GHz

source: https://community.amd.com/community/gaming/blog/2017/03/14/tips-for-building-a-better-amd-ryzen-system

Just checked here. 25 peak consecutive players lol:p

Noone plays this game, literally.

source: http://steamcharts.com/app/228880

It isnt bad if everything else is the same. But everything else isnt. The 1700 is 3 GHz and 7700K is 4.2 GHz. So that is 20% on top of the clock advantage, which is 40% in this particular case.

Overclocking wont change the situation either. You get twice as much chance of getting a Kaby Lake to 5 GHz than a Ryzen to 4 GHz.

As of 3/13/17, the top 29% of 1700Xs were able to hit 4.0GHz or greater.

And remember it was the 1700X, which is a lot more expensive than the 7700K.

As of 2/22/17, the top 59% of tested 7700Ks were able to hit 5.0GHz or greater.

source: https://siliconlottery.com

If you noticed here

Its noticeable if you have 85 Hz monitor and up. Not to mention games will definitely be more demanding in the future. I give a year till the number goes down to 60, which is used by the vast majority of computer users.

source: http://www.techspot.com/review/1348-amd-ryzen-gaming-performance/page5.html

The only thing demanding in the scenario you mentioned is streaming, which can be easily offloaded to the GPU. It is is much more efficient at encoding anyway. That is why AMD themselves released the ReLive driver a couple of months ago, which was a blatant copy of Nvidia’s ShadowPlay that’s been around for years.

The rest can be done even on a $99 smartphone CPU.

Sorry I didnt word it properly, I meant that while additional cores has an advantage, per-core performance is still the most important thing to consider. As @iKirin mentioned, single-core performance is not going away anytime soon. You can only offload so many things to the other core without spending significantly more effort. Programming is hard. Programming for multithreading is a LOT harder. Anyway, workloads that can be spread across many cores like video encoding or 3D rendering, most of the time, is better done on the GPU anyway as it can do that an order of magnitude faster, making the CPU irrelevant in this case as I mentioned in the streaming example above.

In addition to that, major applications may seem to scale well across cores, but may or may not have all its components inside behave that way.

Here’s a good example

I use Autodesk Maya for animation, and after a ton of research, I found that many of the specific processes I work with (which are character rig deformations and viewport playback speed) are single threaded by default.

The proportion is even more staggering once you actually step in the actual world. Many still use older version of the software for cost or compatibility reasons, which, you guessed it, are still single-threaded

source: https://www.reddit.com/r/buildapc/comments/60dcq6/for_those_of_you_considering_buying_a_ryzen_7/

I think comparing the two companies with different cultures, budgeting, strategy, etc. are a bit like comparing apples to oranges. Its a bit like saying “Fourth gen Eve V will have small battery” just because the Surface was done that way.

I personally view Ryzen similarly as the first iteration of the FX, which started is life as the codename Bulldozer. We all know how it ended up.

Well to be fair, i3 has 2 less cores than i5 or i7 and therefore produces way less heat. That is why Intel ships i3 with smaller coolers than its bigger brothers.

Its not just AMD.

Its the common strategy when you don’t have as much R&D budget as your main competitor. You probably have seen the MediaTek SoC that advertises “true 8 core” as opposed to Qualcomm with quad-core or big.LITTLE 8-core configuration. Making a core runs faster is way more difficult, and costly, than just putting more cores.

The benefit is, well, you score higher in benchmarks.

But we all know who provides better overall experience at the end of the day.

(edited to reflect clarifications from AMD's Ryzen is here!)


#86

What intel definitely has is an edge into your wallet.
Sure kabylake is like 20% ipc ahead of ryzen.
this gap will close a little after ryzen has optimised bios and 3rd software got optimised for it.
I will take my hat the next time intel says we have improved the pic by over 50% over the last arch.


#87

Yes, I agree on this :slight_smile: We see a very similar little “war” going on between Qualcomm and Mediatek. I didn’t mean to say that AMD “invented” this strategy, I just wanted to say that they’re using it. Just like Mediatek :slight_smile:

Didn’t know this… Was it always the case or just recently?

And your whole comment… Brilliant :smiley: You expressed everything I wanted to say but didn’t have the words nor concrete information (I’m a lazy ass so I’m lazy to google it out) to say it :smile:


#88

actually i was off 200mhz, but amd agrees on 3.4 ghz https://en.wikipedia.org/wiki/Ryzen
i wrote in my first post that i was talking about the 1700x.

It’s not much played, points for you.

does market share ring a bell ?

here a different bench with actual game names and not just unkown stuff mashed together. bench for 1700c and 1800x http://www.chip.de/artikel/AMD-Ryzen-gegen-Intel-Kaby-Lake-Gaming_111327043.html

Meaning you could set it on mt ?

Did amd also bring 50%ipc boost with bulldozer over the previous arch ?


#89

Ryzen’s IPC in fact is not a lot behind Kaby Lake’s and Skylake’s. It’s just around 5 to 10% depending on type of workload and applications, according to various sources.

What’s lagging behind is clock speed, it’s quite difficult for an eight-core Ryzen to clock higher than 4.2GHz, and it’s not uncommon to see 5GHz Quad-core Kaby Lakes. Comparing stock clocks, the top tier 1800X clocks at 3.6/4.0GHz. The i7-7700K clocks at 4.2/4.5GHz, that’s quite a big difference.

It’s not suprising to see Intel CPUs clock higher, given that Intel’s renown 14nm is state of the art and beats every fab process in the market.


I frankly do not understand why people think Ryzen is the next Bulldozer or that it’s the same thing that happens every year.

At least I personally haven’t seen of anything worth reading about AMD the past few years.
(In hindsight, Excavator is probably worth looking at, even thought its still a bit unimpressive).

Ryzen is different.

Bulldozer performed worse than Phenoms apart from heavily threaded work.
Don’t even bother comparing with Intel’s first and second gen-Core i series. Bulldozer doesn’t beat its own predecessor.

Ryzen beat Bulldozer by 40-50% and caught up with Intel within a competitive margin, even on lower threaded applications. This was something the whole Bulldozer family never even was close of doing.

There’s also nothing to hype about, it’s factual and its truth that AMD would not best Intel any time soon.
It’s just unrealistic to expect AMD making a CPU that’s apparently faster than Intel.
It’s probably way harder for AMD to achieve this than say, making a GPU that beats Nvidia’s.

Intel is a different kind of Goliath and AMD is no David, no game here.

At the bottom line, AMD doesn’t have access to the fab technology Intel does and who is capitalizing upon it.


#90

Sorry, I mean in term of percentage, not the raw number. The difference is actually staggering, as you get 2x as much chance getting 5 GHz on 7700K than getting 4 GHz on 1700X

As of 3/13/17, the top 29% of 1700Xs were able to hit 4.0GHz or greater.

And remember it was the 1700X, which is a lot more expensive than the 7700K.

As of 2/22/17, the top 59% of tested 7700Ks were able to hit 5.0GHz or greater.

Source: https://siliconlottery.com

[quote=“TheDestiny, post:88, topic:5497”]
not just unkown stuff mashed together[/quote]
Sorry I forgot to mention the source. Its http://www.techspot.com/review/1348-amd-ryzen-gaming-performance/page5.html

They compared 22 games and took an average, instead of just a mere 3 games where two of them are sponsored by AMD. Feel free to take a look at the results for each game.

If you noticed, Ryzen took a deep dive in GTA V, which is arguably the most popular title in recent years, almost matching a Core i3 that’s a third of the price. So if you take popularity of each games into account, the difference is even more staggering in my opinion.

Actually I have no idea. I didn’t dive into PC hardware that much back then.

Either way, from the company perspective, Ryzen is closer to FX than anything Intel in term of strategy, budgeting, and R&D resources.

Anyway, we are comparing a new unproven CPU that’s $70 more expensive than the well-proven one. That is on top of the price of the mainboard.

The cheapest Z270 on PCPartPicker is $107. The cheapest X370 is $132, its from Biostar, and its literally the only one that’s less than $150.

Meanwhile, you can get Z270 boards for less than $150 from Gigabyte, MSI, Biostar, ASRock, and Asus, including the ones with SLI support and WiFi. If you want WiFi with X370, you gotta spend $200 on the Asus Taichi.

The difference in overall cost will go over $100 after you took everything into account. I honestly don’t think Ryzen is worth the cost unless youre comparing it against Broadwell-E, which is due for a replacement soon anyway.


#91

the benches i posted are from the 1700, which is as expensive as the i7 7700k.
the 1700 is also able to hit 4 ghz.

yeah real much behind… 4 FPS in ultra settings on 1080p.

and in 1440p in the benches you showed http://www.techspot.com/review/1348-amd-ryzen-gaming-performance/page5.html the 1700 is 13 fps behind the 7700k.

Look for recent benchmarks, not day zero stuff with unoptimised bios.


#93

Ryzen 7 is not targeted towards people who would otherwise choose 7700k.
Ryzen 7 is here to dethrone the expensive 6950k and friends.

Ryzen 5 is the probably the deal for those who’d choose a 7700k.

Yet it’s unreleased, so I’m just talking about the architecture not a specific model nor the price.


#94

I agree, though theyre kind of asking for it, by pricing the 1700 up against the 7700K, and even goes as far as this


#95

So you agree that the 1700 is head on head with the 7700k in 5 of 6 games at day one ?


#96

AMD apparently did. I prefer to believe independent 3rd party testers.


#97

Haha, half of those 3rd party reviews are influenced by intel.
Maybe a few of them give actually objective reviews about ryzen.


#98

Do you have any proof for this?
Given Intel’s history it doesn’t sound too unlikely, but it would be nice to have some way to get to your conclusion by myself :slight_smile:


#99

Not all 1700. In fact, if you got one right now, theres more chance that you actually cannot get 4 GHz than if you can

As of 3/13/17, the top 26% of 1700s were able to hit 4.0GHz or greater.[quote=“TheDestiny, post:91, topic:5497”]
4 FPS in ultra settings on 1080p.
[/quote]

I don’t understand the language, so please correct me if I’m wrong about the following analysis. but I saw that they were testing it with a GTX 960 and 980, which is not a good setup for testing CPU bottlenecking as it shifts the bottleneck from the CPU back to the GPU, especially in Ultra settings. In other words, if you have ‘unlimited’ CPU, the number would be stuck in 40ish fps anyway because the GPU is not able to push more. That is why a Core i5 from 5 years ago can still keep up with the latest and greatest.

On the other hand, as you shift the settings to medium, putting less strain to the GPU, the difference would then get larger.

Well first of all, there is no 1700. There is only the more expensive 1700X in the list. More importantly, 13fps already accounts for more than 10%, which should be noticeable, and can only get larger as time goes on.

Honestly its their fault for rushing the release. The 7700K results are also taken from day zero of its respective release.

I would LOVE to believe that, but as you mentioned…

I too love to make baseless claims, but I think its a bit unfair if I’m allowed to spout BS to make AMD look better than they really are, but not the case with Intel.

On the other hand, and this is one proven, AMD is telling reviewers what to do with the Intel platform. If anything, its AMD the one that has more control over the review results all over the internet.

Some of those tweaks made sense for real-world testing, but others… not so much. One AMD suggestion was indeed disabling the multi-core enhancement/acceleration feature available on many Intel motherboards.

source: https://www.reddit.com/r/hardware/comments/5xb7xb/jayztwocents_says_that_amd_was_told_him_through/dei7aie/


#100

There is no proof i think, those were just rumours about it.
But Intel did it once and got fined for it, was a laughable small fine compared to the loss amd had.
History repeats itself, much likely.


#102

i’m not so much into graphic cards, but do you wanna tell me that a 960/980 can’t handle 1080p ultra settings ?

everything above 60 fps is not noticeable without a monitor with more than 60hz, if you have a better monitor then it may be noticeable.

Those tweaks are in regarding to the ryzen cpu not on how to bench intel cpu’s, you can see benchmarks of the 1700 with set on /off which makes in most games almost zero difference.

j2cents is a special snowflake. watch his latest uploads about ryzen, he makes some interesting and decent content, after putting up an act like a kid when amd told him a day or two before the nda was over to bench again cause there were bios updates LOL.