Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[Zucker2k]

I don't know about you guys but knowing the strengths and weaknesses of an architecture is of interest to some on this forum so let's take both positive and negative news for what they are. Besides, nobody knows as yet what the actual performance of the chips are so we're all speculating. Personally, I'm peeved by the wild, overoptimistic speculations based on scanty, unverified information. This only serves to dampen what is an excellent product with an excellent price tags from AMD should the reviews fail to meet the high expectations set by a few. I've seen this trend getting repeated with every AMD cpu release since 2006. Please, hold your horses and be a little more tolerant of the views and opinions of other people, for the reviews are less than 24 hours away. We shall know soon enough.

Something I just noticed about the leaked gaming slides from AMD:

There are some serious gpu bottlenecks in there. And why is AMD using a less powerful gpu for the R7 1700 and the i7 7700k? It's clear both those chips could stretch their legs if not constrained by the gtx 1070 at those resolutions. Three theories I have about the above:
1. mask the little to no difference between the R7 1700 and R7 1700X
2. hold back the competitor chip from showing it's strength
3. both 1 and 2

 

[rbk123]

Since you guys do not understand, or maybe you do it on purpose i will try to explain what those benchmarks means...

Let's say, for arguments sake, you are correct and AMD's product is another Bulldozer. What do you get as your prize for that? You get to pay more for the same chips you want from Intel instead of less. That's certainly reason to celebrate.

You aren't very bright are you?

 

[Arachnotronic]

Nope,but he said it was Bulldozer again - Bulldozer regressed in IPC.

No, Ryzen is not Bulldozer, not by a long shot. It appears to be a well designed, well balanced, and power efficient architecture. Bulldozer was none of those things.

 

[tamz_msc]

yeah right...from Amd slides...how do you know they both were running at 4ghz?

As i said in my previous post if Ryzen has same IPC as broadwell why does it loses on games? It is also running at higher frequency..

Again, you don't even bother to try googling do you?

 

[USER8000]

No, Ryzen is not Bulldozer, not by a long shot. It appears to be a well designed, well balanced, and power efficient architecture. Bulldozer was none of those things.

Yep,and even on UK forums,many people didn't expect AMD to be better than Broadwell IPC ,maybe between Haswell and Broadwell IPC at best,and not to hit the same clockspeeds as Kaby Lake.

OTH,many of us see AMD offering good enough single core IPC and decent core counts for much less money than Intel as a good thing,as quad cores and six core CPUs will be far cheaper than the £330+ a Core i7 7700K costs now. Most gamers are GPU limited,so the money save goes on a better graphics card.

 

[PhonakV30]

https://videocardz.com/66640/amd-ryzen-7-launches-tomorrow

 

[Arachnotronic]

Yep,o and even on UK forums,many people didn't expect AMD to be better than Broadwell IPC ,maybe between Haswell and Broadwell IPC at best,and not to hit the same clockspeeds as Kaby Lake.

OTH,many of us see AMD offering good enough single core IPC and decent core counts for much less money than Intel as a good thing,as quad cores and six core CPUs will be far cheaper than the £330+ a Core i7 7700K costs now. Most gamers are GPU limited,so the money save goes on a better graphics card.

AMD's pricing is really the star of this show -- they sensed that enthusiasts wanted "more cores" and wanted those cores to be decent and not priced to the moon, and they appear to have delivered.

What's surprising to me is how Intel has not yet cut prices on Broadwell-E. They have to know that nobody is going to want to buy those now, except maybe the top 10 core.

 

[RoarTiger]

So AMD is now the power efficient company? What is happening!

 

[USER8000]

AMD's pricing is really the star of this show -- they sensed that enthusiasts wanted "more cores" and wanted those cores to be decent and not priced to the moon, and they appear to have delivered.

What's surprising to me is how Intel has not yet cut prices on Broadwell-E. They have to know that nobody is going to want to buy those now, except maybe the top 10 core.

Yep,and lower core count models will be far cheaper. Its one thing have a 60% to 70% single core performance deficit - but once you get to within 20% to 25% against the top models it starts to become more diminishing returns,and is why Sandy Bridge has lasted so long.

OTH,most of the Intel Core i5 range is dominated by lower clockspeed models like the Core i5 7400,etc which are in more danger than a Core i7 7700K.

With the huge reductions in power consumption,the APUs have a real chance now in laptops,which is one area Intel and Nvidia are destroying AMD in.

 

[imported_jjj]

I don't know about you guys but knowing the strengths and weaknesses of an architecture is of interest to some on this forum so let's take both positive and negative news for what they are. Besides, nobody knows as yet what the actual performance of the chips are so we're all speculating. Personally, I'm peeved by the wild, overoptimistic speculations based on scanty, unverified information. This only serves to dampen what is an excellent product with an excellent price tags from AMD should the reviews fail to meet the high expectations set by a few. I've seen this trend getting repeated with every AMD cpu release since 2006. Please, hold your horses and be a little more tolerant of the views and opinions of other people, for the reviews are less than 24 hours away. We shall know soon enough.

Something I just noticed about the leaked gaming slides from AMD:

There are some serious gpu bottlenecks in there. And why is AMD using a less powerful gpu for the R7 1700 and the i7 7700k? It's clear both those chips could stretch their legs if not constrained by the gtx 1070 at those resolutions. Three theories I have about the above:
1. mask the little to no difference between the R7 1700 and R7 1700X
2. hold back the competitor chip from showing it's strength
3. both 1 and 2

The entire point of gaming is to be GPU bound. If you test at unrealistic resolutions, you aren't testing gaming anymore.
Just the fact that someone thinks that you need a high end CPU for gaming, is so utterly sad and shows how much damage Intel has done.
Claiming to be testing gaming when at low res is like testing a CPU with 5MB of RAM, you create a bottleneck that doesn't exist and you get results of no relevance.
Testing at slightly lower res for the GPU is fine if you want to look at how future proof the CPU is but that only works if you assume that games won't scale to more cores ( and that resolution remains the same) and that assumption is false at this point.

 

[USER8000]

So AMD is now the power efficient company? What is happening!

Just wow - this might be why there are rumours of one of the laptop OEMs putting a Ryzen 7 in a desktop replacement notebook.

 

[formulav8]

Were you expecting Ryzen to beat Intels latest and greatest in ST performance?

Yeah I can't believe people are actually thinking/expecting that. AMD has caught up quite well in ST, but some people expect the impossible, even with the budget AMD had to work with.

I pay more attention to the price/performance metric when I buy something, not simply performance by itself. And if I need a new system upgrade, (like I do now whether I buy AMD or Intel next), I look at the total platform cost.

 

[Zucker2k]

The entire point of gaming is to be GPU bound. If you test at unrealistic resolutions, you aren't testing gaming anymore.
Just the fact that someone thinks that you need a high end CPU for gaming, is so utterly sad and shows how much damage Intel has done.
Claiming to be testing gaming when at low res is like testing a CPU with 5MB of RAM, you create a bottleneck that doesn't exist and you get results of no relevance.
Testing at slightly lower res for the GPU is fine if you want to look at how future proof the CPU is but that only works if you assume that games won't scale to more cores ( and that resolution remains the same) and that assumption is false at this point.

I'm not sure if you're serious. Would you play "Ashes of the Singularity" at 30fps and lower? The bottleneck here is artificial, and an unforeseen consequence of AMD's alleged decision-making. The test is not a gpu test; let the cpu carry the load. It is not a coincidence that the faster gpu delivered more frames so why attempt to hide that fact?

 

[plopke]

I am trying really hard to follow this conversation , so people are disappointed in these gaming numbers and surprised that the 7700K might still be overall the best gaming cpu? Isn't that what we always expected , but for me personally I would go for example for a 1700 over a 7700K because of my slow upgrade cycles of my cpu. If you tend to upgrade your cpu a lot quicker and are a extreme gamer , just get a 7700K no need to get upset ?

 

[.vodka]

So AMD is now the power efficient company? What is happening!

Not surprising. Expected, actually.

It's a dual channel platform vs a quad channel platform
Most "southbridge" functions integrated in the CPU vs a more power hungry X99 chipset
Smaller, leaner, power efficient cores without 256 bit execution units and required power hungry structures to support that, vs overbuilt cores with overbuilt uncore
etc

In some workloads Intel's cores will have the lead thanks to full rate AVX/2 throughput and therefore justifying the power consumption... but in the other 95% of use cases, it'll probably be like this.

Could you post the rest of this? It's a wccftech URL but I can't find the article.

Ryzen has just found its way in servers and the datacenter with the performance it has, and this power consumption if it holds up in these workloads.

 

[USER8000]

I'm not sure if you're serious. Would you play "Ashes of the Singularity" at 30fps and lower? The bottleneck here is artificial, and an unforeseen consequence of AMD's alleged decision-making. The test is not a gpu test; let the cpu carry the load. It is not a coincidence that the faster gpu delivered more frames so why attempt to hide that fact?

You might actually want to look at more RTS games then - games like Supreme Commander and Sins of a Solar Empire during late game battles could easily have under 30FPS,and even employ time dilation.

Most people are not running a GTX1070 or a similar performance card anyway.

 

[formulav8]

AMD Ryzen motherboards explained: The crucial differences in every AM4 chipset

 

[RoarTiger]

Could you post the rest of this? It's a wccftech URL but I can't find the article.

From todays article linked earlier in this thread. Middle of page.
http://wccftech.com/amd-ryzen-7-1700x-1700-official-gaming-benchmarks-leak/

 

[coercitiv]

I'm not sure if you're serious. Would you play "Ashes of the Singularity" at 30fps and lower?

The short answer is YES. The exact answer includes words such as RTS and 99th percentile. The long answer also includes details about AOTS and how any GPU intensive load in that game is also very CPU intensive.

Less than 24 hours until gaming benches at 720p/1080p tell the full story.

 

[USER8000]

First review leaked:

http://www.shahrsakhtafzar.com/fa/review/cpu/11652-amd-ryzen-7-1700x-review

A poster on Overclock.net made the following comment:

I think it's important to note the ff (as far as I can tell)

-ram running only at 2133mhz
-cheapo cooler used so no XFR
-they manage to oc it at 4ghz but didn't run the clock on mosts test

 

[unseenmorbidity]

Wasn't he only a few weeks ago participating vigorously in threads on his forum, suggesting that Zen was overrated and/or was quite an unbalanced design?

Now he is publicly singing the praises of Zen?

Life comes at you fast.

Money is a hell of a drug aint it?

At least we know to instantly dismiss his opinion from now on.

 

[zinfamous]

I don't know why you are making fun of anybody who said that Ryzen had Sandy Bridge IPC. Based on AMD's 40% claim, that's where it would have landed. There is no way for anybody who was making predictions before these announcements to know that AMD was sandbagging and actually got 52%.

But this is March 2017, not March 2016. There is a vast array of information available now, freely, to everyone, that is a clear update to what was claimed an entire year ago.

making fun of someone claiming that RyZen is SB IPC today, March 2017 and despite an entire year's worth of updated information, official tests and highly plausible leaked benchmarks that are easily available to that person, should be the standard. It should be lauded. Such people aren't worthy of much consideration, imo.

AMD announced as long ago as November (or October?) that they had, in fact, beaten their 40% projected gain, and quite significantly at that. Why was that announcement not as official as their earlier claim of "only 40%?" Or, do some users just like to ignore inconvenient information for, well, no real reason?

 

[formulav8]

AMD sent reviewers 3000 MHz ram, why would they run it at 2133?

 

[USER8000]

AMD sent reviewers 3000 MHz ram, why would they run it at 2133?

Look at the overclock.net comment - its most likely the reviewer bought the hardware themselves.

Power consumption is very low - so I think there must be a hard 95W TDP limited there.

 

[DarkKnightDude]

I'm annoyed at that Iranian leaked review honestly. Paired it only with 2133mhz RAM and used resolutions like 1366x768? 1600x900? If you've got that much money to drop on a mobo and that cpu surely you have a 1080p monitor at least. Tetsing ita t that resolution is arbitrary at best.

The memory latency thing though seems to be high like every leak says. What is up.