Ryzen "2" ?

humbug · 2 Apr 2018 at 13:17

Ok watched it

There might actually be an IPC uplift here.

Cinebench: 1800X
MT: 1782 @ 4.1Ghz
ST: 167 @ 4.1Ghz

Cinebench: 2700X
MT: 1948 @ 4.3Ghz (+9.3%)
ST: 178 @ 4.3Ghz (+6.6%)

4.1Ghz to 4.3Ghz is a 4.9% difference, take that off the differences is scores and you're left with +4.4% MT IPC and a +1.7% Difference in ST IPC.

Reference https://www.kitguru.net/components/cpu/luke-hill/amd-ryzen-7-1800x-cpu-review/5/

kitfit1 · 2 Apr 2018 at 13:20

Just ran this on my 1700 as was thinking the same thing as you humbug. No need for me to work out the uplift now though

humbug · 2 Apr 2018 at 13:25

kitfit1 said:
Just ran this on my 1700 as was thinking the same thing as you humbug. No need for me to work out the uplift now though

3.8Ghz: 158
4.3Ghz: 178 (+12.6%)

3.8Ghz vs 4.3Ghz = 13% (-0.4% to your 1700) so no, perhaps not on the ST

kitfit1 · 2 Apr 2018 at 13:50

humbug said:
3.8Ghz: 158
4.3Ghz: 178 (+12.6%)

3.8Ghz vs 4.3Ghz = 13% (-0.4% to your 1700) so no, perhaps not on the ST

Yep, sorry to be the bearer of bad news. Well, Gavin will be back in here as soon as he sees that.........................like a rat up a drainpipe

LOLN1 · 2 Apr 2018 at 14:14

Hello
I'm at lunch with the family I follow you via smartphone
My tests are completely provisional;
There is no new bios gigabyte's for 2700x
In the bios some parameters have no effect on the CPU
Ryzen Master does not recognize the cpu
more experienced people than me will do better tests
My conclusions are:
Big CPU at default (only problem too high voltage default)
In overclocking with custom heatsink we can have a lot of power on the CPU
we need new bios and new Ryzen Master
Non ci sono nuovi bios gigabyte per 2700x
Nel bios alcuni parametri non hanno alcun effetto sulla CPU
Ryzen Master non Intel la CPU
Le persone più esperte di me faranno il test
Le mie conclusioni sono:
Grande CPU di default (solo problema troppo alta tensione predefinita)
Nell'overclocking con dissipatore di calore personalizzato potrai avere molta potenza sulla CPU
abbiamo bisogno di nuovo bios e nuovo Ryzen Master

humbug · 2 Apr 2018 at 14:18

kitfit1 said:
Yep, sorry to be the bearer of bad news. Well, Gavin will be back in here as soon as he sees that.........................like a rat up a drainpipe

LOLN1 said:
Hello
I'm at lunch with the family I follow you via smartphone
My tests are completely provisional;
There is no new bios gigabyte's for 2700x
In the bios some parameters have no effect on the CPU
Ryzen Master does not recognize the cpu
more experienced people than me will do better tests
My conclusions are:
Big CPU at default (only problem too high voltage default)
In overclocking with custom heatsink we can have a lot of power on the CPU
we need new bios and new Ryzen Master
Non ci sono nuovi bios gigabyte per 2700x
Nel bios alcuni parametri non hanno alcun effetto sulla CPU
Ryzen Master non Intel la CPU
Le persone più esperte di me faranno il test
Le mie conclusioni sono:
Grande CPU di default (solo problema troppo alta tensione predefinita)
Nell'overclocking con dissipatore di calore personalizzato potrai avere molta potenza sulla CPU
abbiamo bisogno di nuovo bios e nuovo Ryzen Master

Thanks, good job

NinjaCool · 2 Apr 2018 at 14:19

The Cinebench scores of the 2700X in the live stream is not bad at all

It will be interesting to see how much we can push it on the new X470 boards.

kitfit1 · 2 Apr 2018 at 14:24

LOLN1 said:
Hello
I'm at lunch with the family I follow you via smartphone
My tests are completely provisional;
There is no new bios gigabyte's for 2700x
In the bios some parameters have no effect on the CPU
Ryzen Master does not recognize the cpu
more experienced people than me will do better tests
My conclusions are:
Big CPU at default (only problem too high voltage default)
In overclocking with custom heatsink we can have a lot of power on the CPU
we need new bios and new Ryzen Master
Non ci sono nuovi bios gigabyte per 2700x
Nel bios alcuni parametri non hanno alcun effetto sulla CPU
Ryzen Master non Intel la CPU
Le persone più esperte di me faranno il test
Le mie conclusioni sono:
Grande CPU di default (solo problema troppo alta tensione predefinita)
Nell'overclocking con dissipatore di calore personalizzato potrai avere molta potenza sulla CPU
abbiamo bisogno di nuovo bios e nuovo Ryzen Master

I think we all know those results were provisional LOLN1, for all we know that might have been the first time you had even run the 2700x ?
Either way, a big thanks for giving all of us a little preview of what to expect. Personally i will defo get a 2700x on release, even though at the moment there seems to be no IPC increase. As a non gamer, more of a bencher myself..........................it's about benching against other people with the same hardware.

kitfit1 · 2 Apr 2018 at 14:30

NinjaCool said:
The Cinebench scores of the 2700X in the live stream not bad at all
It will be interesting to see how much we can push it on the new X470 boards.

As you can see from my post above though, the 2700x result on the face of it looks good. But compared against an existing 1700 clock for clock, it's actually nothing to write home about.
I'm putting it down to the bios LOLN1 was running on his Gigabyte mobo more than anything. Just wished he had used a VI Hero as the latest bios is optimized for the Zen refresh already.

humbug · 2 Apr 2018 at 14:39

If these are already with some retailers it can't be long now, also, they launch on the 19'th.

So far i'm 'while not bowled over' also not disappointed, Ryzen 1### was always an impressive workhorse anyway so 10% added on to that is great.

The compatible 300 series Motherboard BIOS and 400 series Motherboard testing will be interesting given what CanardPC said about the gaming performance gain they got with their X470.

DragonQ · 2 Apr 2018 at 14:40

humbug said:
Ok watched it

There might actually be an IPC uplift here.

Cinebench: 1800X
MT: 1782 @ 4.1Ghz
ST: 167 @ 4.1Ghz

Cinebench: 2700X
MT: 1948 @ 4.3Ghz (+9.3%)
ST: 178 @ 4.3Ghz (+6.6%)

4.1Ghz to 4.3Ghz is a 4.9% difference, take that off the differences is scores and you're left with +4.4% MT IPC and a +1.7% Difference in ST IPC.

Reference https://www.kitguru.net/components/cpu/luke-hill/amd-ryzen-7-1800x-cpu-review/5/

humbug said:
3.8Ghz: 158
4.3Ghz: 178 (+12.6%)

3.8Ghz vs 4.3Ghz = 13% (-0.4% to your 1700) so no, perhaps not on the ST

As I've told you before, IPC and clock speed changes are multiplicative not additive. You can't just say "6.6% overall improvement - 4.9% clock speed bump = 1.7% IPC improvement". If you have 10% IPC improvement and 10% clock speed bump, for example, you get a 1.1 x 1.1 = 1.21x improvement, not 1.1 + 1.1 = 1.2x.

In this case You need to adjust 178 score to 4.1 GHz (assuming linear clock speed) to get 169.7. Then you can see that's 1.6% better than a score of 167. However, that is well within the margin of error and RAM speed/timing differences could easily account for this. I doubt an IPC increase for Cinebench.

humbug · 2 Apr 2018 at 14:46

DragonQ said:
As I've told you before, IPC and clock speed changes are multiplicative not additive. You can't just say 6.6% - 4.9% = 1.7% IPC. Adjust 178 score to 4.1 GHz (assuming linear clock speed) to get 169.7. Then you can see that's 1.6% better than a score of 167.

However, that is well within the margin of error and RAM speed differences could easily account for this. I doubt an IPC increase for Cinebench.

You're stating the obvious, Its impossible to account for scaling or margins of error, difference in reference data ecte.... its why i said "Might" if you didn't read that or understand what "Might" means then that's your problem, not mine, only you are over reacting to what was clearly an exercise in speculation. If you're going to be that much of a Nazi over it then its you who have issues.

orbitalwalsh · 2 Apr 2018 at 14:48

kitfit1 said:
As you can see from my post above though, the 2700x result on the face of it looks good. But compared against an existing 1700 clock for clock, it's actually nothing to write home about.
I'm putting it down to the bios LOLN1 was running on his Gigabyte mobo more than anything. Just wished he had used a VI Hero as the latest bios is optimized for the Zen refresh already.

already has bridge bios

https://www.gigabyte.com/us/Motherboard/GA-AX370-Gaming-5-rev-10#support-dl
F22
Update AGESA 1.0.0.1a

https://www.asus.com/uk/Motherboards/ROG-CROSSHAIR-VI-HERO/HelpDesk_Download/

asus is still on 1.0.0.0a

Version 3502 2018/01/

CROSSHAIR VI HERO BIOS 3502
Update to AGESA 1000a for new upcoming processors

1.0.0.1a get the CPU supported/working nothing about full optimisations etc

ASRock also has access to 1.0.0.1a

http://www.asrock.com/mb/AMD/X370 Taichi/index.asp#BIOS
Update AMD AGESA to PinnaclePI-AM4_1.0.0.1a.

MSI also has it rolled out

kitfit1 · 2 Apr 2018 at 15:07

orbitalwalsh said:
already has bridge bios

https://www.gigabyte.com/us/Motherboard/GA-AX370-Gaming-5-rev-10#support-dl
F22
Update AGESA 1.0.0.1a

https://www.asus.com/uk/Motherboards/ROG-CROSSHAIR-VI-HERO/HelpDesk_Download/

asus is still on 1.0.0.0a

Version 3502 2018/01/

CROSSHAIR VI HERO BIOS 3502
Update to AGESA 1000a for new upcoming processors

1.0.0.1a get the CPU supported/working nothing about full optimisations etc

ASRock also has access to 1.0.0.1a

http://www.asrock.com/mb/AMD/X370 Taichi/index.asp#BIOS
Update AMD AGESA to PinnaclePI-AM4_1.0.0.1a.

MSI also has it rolled out

6001 for the VI Hero, is still in Beta. Maybe Asus thinks it needs more testing yet. Obviously LOLN1 was having issues with bios for his Gigabyte mobo as he has said himself and was obvious during the Livestream..............................or havn't you bothered watching it :rolleyes:

orbitalwalsh · 2 Apr 2018 at 15:26

kitfit1 said:
6001 for the VI Hero, is still in Beta. Maybe Asus thinks it needs more testing yet. Obviously LOLN1 was having issues with bios for his Gigabyte mobo as he has said himself and was obvious during the Livestream..............................or havn't you bothered watching it

Asus is the same 1.0.0.1a code

All will be the same till next bios release for all vendors .

key thing is shows 1.0.0.1a works! and does its job- rzen 2000 series an work off the bat. again optimisations will follow launch day (hopefully) , date should be floating about by now

DragonQ · 2 Apr 2018 at 16:10

humbug said:
You're stating the obvious, Its impossible to account for scaling or margins of error, difference in reference data ecte.... its why i said "Might" if you didn't read that or understand what "Might" means then that's your problem, not mine, only you are over reacting to what was clearly an exercise in speculation. If you're going to be that much of a Nazi over it then its you who have issues.

I feel like it's right to correct you because you're stating things as fact again when they aren't facts.

humbug · 2 Apr 2018 at 16:16

DragonQ said:
I feel like it's right to correct you because you're stating things as fact again when they aren't facts.

I which case because of your trolling eagerness to do that you ###### up, because i didn't, i said "it Might"
Its all in your head, if you have an irrational preconception about something you will see it even when its not there, you cannot see what was clearly speculation for your need to make the evidence fit your preconceptions, or some troll is using you. so get a grip, no conspiracy here.

DragonQ · 2 Apr 2018 at 16:32

humbug said:
I which case because of your trolling eagerness to do that you ###### up, because i didn't, i said "it Might"
Its all in your head, if you have an irrational preconception about something you will see it even when its not there, you cannot see what was clearly speculation for your need to make the evidence fit your preconceptions, or some troll is using you. so get a grip, no conspiracy here.

OK bro, no need to be so melodramatic and defensive. I don't remember mentioning anything about a conspiracy.

humbug · 2 Apr 2018 at 16:38

DragonQ said:
OK bro, no need to be so melodramatic and defensive. I don't remember mentioning anything about a conspiracy.

These arguments are a pain in the arse, i ignored the first 'correction' so not to escalate it, Yes i know on that occasion you were right but it was petty, it seems to have escalated anyway, let us just agree to quit, the last thing this forum needs is more ongoing feuds.

kitfit1 · 2 Apr 2018 at 16:44

humbug said:
These arguments are a pain in the arse, i ignored the first 'correction' so not to escalate it, Yes i know on that occasion you were right but it was petty, it seems to have escalated anyway, let us just agree to quit, the last thing this forum needs is more ongoing feuds.

Yep.......................come on girls.................close the handbags, bye each other a Gin and Tonic and a packet of scratchings