Ryzen 2 improves by 10%?!?

[BluJay614]

https://segmentnext.com/2018/01/23/amd-ryzen-5-2600-performs/

Well, we all have heard about the second generation of Ryzen processors. The question is, how much will they really perform?

While we don't have a way to verify this information, apparently a leak has shown up demonstrating a 3.3-10.1% improvement with the 2600 over the 1600 depending on the type of workload being done. Not by a ton, and there is no way to verify this, but all the same, if this was even done with a early version that still wasn't final, it could still get better from there.

[NTDaws]

Better than 0%

[Ryujin2003]

I thought we've already posted this?

[Energycore]

Please edit your thread to meet our Posting Guidelines.

 

Once you have, mention me @Energycore before 8h have passed and I'll move it back

.

[aezakmi]

FX 2

[mynameisjuan]

Im pretty sure 10% was predicted a month after Ryzen was released as the architecture is already pushed close to its limits. 

[The Benjamins]

1 minute ago, Being Delirious said:

Better than 0%

If we take the clocks and scores accurately, this means it has a 5% clock bump and a 9% performance bump, meaning we are seeing a 4~% IPC bump.

[LeapFrogMasterRace]

gibe me more clocks plz

[onlybuilt4cubanxlinx]

Just hoping for better launch than Ryzen with out major bugs and I'll finally upgrade from a 8320e to one. 

[Mira Yurizaki]

If the IPC improvement is 10% or so, then that's what I honestly expect out of anyone. I think it's practically a myth that IPC improvements have been anymore than say 10-15% on average except in two cases.

 

Most of the improvements we've been enjoying were from clock speed bumps.

[xg32]

16 minutes ago, BluJay614 said:

https://segmentnext.com/2018/01/23/amd-ryzen-5-2600-performs/

Well, we all have heard about the second generation of Ryzen processors. The question is, how much will they really perform?

While we don't have a way to verify this information, apparently a leak has shown up demonstrating a 3.3-10.1% improvement with the 2600 over the 1600 depending on the type of workload being done. Not by a ton, and there is no way to verify this, but all the same, if this was even done with a early version that still wasn't final, it could still get better from there.

5% is expected, 10% is huge (imagine 6 and 8 cores running at 4.4), it's already confirmed that it can run 3600-3800 on the ram and we know what that does for ryzen.

[NinJake]

The real cool thing is if you are in the market for building a new pc in the upcoming future, you can take that ~10% performance increase in the Ryzen 2 chips... but the 1st gen Ryzen chips will hopefully drop quite significantly in price >10% price cut which will save you more money that you can put elsewhere.

 

Of course once the new of the new is out, some people don't settle for anything less than the "best" or the "latest and greatest". I'm still excited, as I'm planning on putting together a new build sometime this year or early 2019. All I know is it will be based off of the 1180Ti once it's released... hopefully that will see more than a 10% performance increase compared to the 980Ti (I kid, I'm certain it will.)

[Coaxialgamer]

This is still based on the Zen 1 core . I wouldn't expect any real IPC gain. That said , memory controller might bring performance gains. And the 10% clock speed boost from 12nm is also nice.

[Stefan Payne]

6 hours ago, BluJay614 said:

https://segmentnext.com/2018/01/23/amd-ryzen-5-2600-performs/

Well, we all have heard about the second generation of Ryzen processors. The question is, how much will they really perform?

While we don't have a way to verify this information, apparently a leak has shown up demonstrating a 3.3-10.1% improvement with the 2600 over the 1600 depending on the type of workload being done. Not by a ton, and there is no way to verify this, but all the same, if this was even done with a early version that still wasn't final, it could still get better from there.

Depends on how much they can optimize the Memory Controller and get the Latency down.

There are some things you can do...

[i_build_nanosuits]

6 hours ago, M.Yurizaki said:

If the IPC improvement is 10% or so, then that's what I honestly expect out of anyone. I think it's practically a myth that IPC improvements have been anymore than say 10-15% on average except in two cases.

 

Most of the improvements we've been enjoying were from clock speed bumps.

IPC improvement is confirmed at 0%

the 10% increase is in clock speed...so slightly better overclocks...maybe.

[rukspuks]

10% is pretty good if the prices remain the same. Intel chips too gain 5-10% per generation if all other variables are equal.

[Ebony Falcon]

We don't no anything yet apart from one keeled engineer same which could have been from 5 months ago 

 

maxwell and pascal are the same architecture on a node shrink look how much that gained 

 

there will be a small ipc gain from infinity fabric being tuned and clocks out of the box should be 10 percent faster but we don't no anything else about the silicon other that it's built from the ground up for higher clock speeds at the sacrifice of yields 

 

10 percent out of the box isn't a lot but what if they oc like a beast ? We just don't know untill we get them

 

ryzen 1 was worst case scenarios for amd it's what they said, ryzen 1 was the worst possible case for the cpu and it's still great

 

ryzen plus could be fantastic and oc beyond belief no one knows 

 

10 percent could turn into 20 with oc

 

8 cores balling at 4.8ghz would destroy Intel but we just don't no 

[porina]

8 hours ago, rukspuks said:

10% is pretty good if the prices remain the same. Intel chips too gain 5-10% per generation if all other variables are equal.

For practical purposes IPC hasn't moved on Intel side since Skylake, outside of some specific new instructions.

 

To clarify here, I'd break down IPC into two categories, peak IPC, and system IPC. Peak IPC is what is the CPU capable of if not hitting external bottlenecks. This is easier to test for, with small code that fits within on CPU caches. System IPC then does start including more factors, and makes more use of cache size/speed, as well as ram configuration. As example, for general consumer tasks, Skylake and Skylake-X have same peak IPC (outside of AVX-512 anyway), but Skylake-X has lower system IPC in some workloads due to the new L3 cache structure, but could be higher system IPC in others due to more ram bandwidth.

 

Ryzen 2nd gen could be a similar parallel here, with same peak IPC as cores aren't believed to be changed significantly, but system IPC may increase with optimisations in cache and ram access. It should help some things, but doesn't necessarily help everything.

 

And that's before we look at potential clock gains...