Ryzen "2" ?

HeX · 15 May 2018 at 17:39

southernorth said:
That's Ryzen+ already released

Nope, it's Ryzen 2 / Zen+, next year will be Ryzen 3 / Zen 2.

Ryzen is the brand, Zen is the architecture.

Zeed · 15 May 2018 at 18:23

Doobedoo said:
Was that back when they got caught rigging the market? Or was that also due to mobile phone demand too?

I could actually understand if their was an actual shortage and their profit margins hadn't increased as much as ram pricing, as it stands I could quite easily buy 20 kits if I wanted to so ram is clearly available, unlike the recent gpu mining craze.

DRAM makers sued (yet again) for 'fixing prices' (yet again) of chips https://www.theregister.co.uk/2018/04/30/dram_vendors_sued_again_for_price_fixing_again/

They can get sued as every freaki g week pay fee and **** on i t all with bilions they are making.

Have you seen GN video about ram prices and why they will never go down??

Fantastic point in the video was that they have not even jnew WHO CORSAIR IS thats how much money they make on server market they give 0 **** about desktop market. Simple.

opethdisciple · 15 May 2018 at 20:02

I received my RMA of my ram.

But the box was not sealed. Now I'm now unsure if OcUK just sent it back to me without switching the ram.

I'll try it. Fingers crossed it doesn't freeze on me.

Gee · 15 May 2018 at 20:06

No note in the box?

opethdisciple · 15 May 2018 at 20:13

Gee said:
No note in the box?

You mean the little warranty leaflet? Yea that's there.

Potty27 · 15 May 2018 at 20:36

The 8 Pack RAM I bought new was also not sealed properly... but it works so I'm happy with it, weird though.

chroniclard · 15 May 2018 at 20:49

prjwebb said:
Timespy has Graphics and CPU tests, no combined. It's the combined in Firestrike that Ryzen is horrendous in so you can't say it's better in Timespy because there's literally no test for it.

Horrendous? Not the right word.

sandys · 15 May 2018 at 21:16

prjwebb said:
Timespy has Graphics and CPU tests, no combined. It's the combined in Firestrike that Ryzen is horrendous in so you can't say it's better in Timespy because there's literally no test for it.

My Ryzen seems OK think its just the amount of memory tweaking and tuning you have to do to get the performance out of it.

opethdisciple · 15 May 2018 at 21:29

The RMA'd ram is still freezing.

Everything is stock other than DOCP.

I'll now try it at 3133MHz.

Pc seems much happier at 3133MHz.

prjwebb · 15 May 2018 at 22:01

sandys said:
My Ryzen seems OK think its just the amount of memory tweaking and tuning you have to do to get the performance out of it.

My 8600k would get over 12k if I pushed the oc. That's just a 6c

sandys · 15 May 2018 at 22:25

opethdisciple said:
The RMA'd ram is still freezing.

Everything is stock other than DOCP.

I'll now try it at 3133MHz.

Pc seems much happier at 3133MHz.

I can't run XMP timings on mine, you probably need to forget those altogether if the RAM is not qualified on your board.

prjwebb · 15 May 2018 at 22:36

sandys said:
My Ryzen seems OK think its just the amount of memory tweaking and tuning you have to do to get the performance out of it.

But they is much better than 1000 series

opethdisciple · 15 May 2018 at 22:37

sandys said:
I can't run XMP timings on mine, you probably need to forget those altogether if the RAM is not qualified on your board.

You have 8 pack B-Die ram as well?

Shame really as I was lead to believe this 8 Pack ram would solve all my issues.

So far it seems to be running 3133Mhz just fine. So as long as that stays true. It'll do.

sandys · 15 May 2018 at 22:44

prjwebb said:
My 8600k would get over 12k if I pushed the oc. That's just a 6c

Regardless, looks like Firestrike as not been updated to take account of so many threads on the combined test, developer acknowledges this, suggests you use Time Spy Extreme, so no drama here.

sandys · 15 May 2018 at 22:46

opethdisciple said:
You have 8 pack B-Die ram as well?

Shame really as I was lead to believe this 8 Pack ram would solve all my issues.

So far it seems to be running 3133Mhz just fine. So as long as that stays true. It'll do.

I don't have 8 pack RAM, some fancy pants Trident Z RGB that wouldn't run at any of its rated XMP settings, so I feel your pain. Took me 2 weeks of tweaking and testing to get stuff sorted, basically if 8 pack is not on motherboard QVL your are **** out of luck, you need to do the work, or refund and get something on QVL list.

There is a list of B Die stable timings for Ryzen on another board from the Stilt or soething like that, you will likely have better success starting there.

BongoHunter · 15 May 2018 at 23:14

Anyone else think DDR4 prices are coming down in price a bit?

32GB (2x16) of Team Group Vulcan down to about £250 at the moment, still not fallen back to where it was (got my 64GB for £350) - but seems to be getting a little cheaper compared to late last year at least

chroniclard · 15 May 2018 at 23:19

Time spy comparison, not a lot in it in this specific benchmark.

8700k @ 5ghz, same GPU with similar clocks.
2700X is "only" doing 4161 all core at this point but does 4445 boost single core.

Might try with an all core overclock instead of boost, see what effect it has.

gavinh87 · 15 May 2018 at 23:27

chroniclard said:
Time spy comparison, not a lot in it in this specific benchmark.

8700k @ 5ghz, same GPU with similar clocks.
2700X is "only" doing 4161 all core at this point but does 4445 boost single core.

Might try with an all core overclock instead of boost, see what effect it has.

Hardly surprising there's little in it. That's a 1080 doing 4k lol. You could probably throw an atom into the mix and get similar results

prjwebb · 15 May 2018 at 23:46

sandys said:
Regardless, looks like Firestrike as not been updated to take account of so many threads on the combined test, developer acknowledges this, suggests you use Time Spy Extreme, so no drama here.

The fact that Ryzen 2 does so much better than Ryzen 1 in the combined suggests it's a latency issue to me, as there's still gains with cores across the same platform.

But yeah, it's much easier to use a 4k benchmark and say they are close.

sandys · 16 May 2018 at 00:06

The developer suggests the combined score issue is one thing and suggests you use a later variant of their benchmark, you some how know better than the dev?

The simulation at the end of TSE is CPU based not GPU, 4k doesn't come into it.
Regarding TSE

TSE with rx480 http://www.3dmark.com/spy/2580382
TSE with 1080TI https://www.3dmark.com/spy/2593126

Note the negligible difference in CPU score because despite crap GPU the CPU test is only testing CPU and so unaffected by GPU and res!

Anyway all unimportant my system kick arse big time even if it is a slow arse Ryzen