Soldato
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Zen = Sandybridge IPC
- Thread starter RavenXXX2
- Start date
Associate
- Joined
- 9 Jan 2019
- Posts
- 885
******** to that lol
I may be somewhat missing the point, and tbh I don't really care about gaming anyway, but why test 4c8t Sandy vs 4c8t Zen? Yes he's showing that core for core they're similar (swings and roundabouts between games), but that's not taking account of the fact that Zen is actually 8c16t. What do his benchmarks look like after he runs the CPUs as intended? It would have been interesting to see some server and workstation tasks in amongst the benchmarks too.
Soldato
That doesn't seem right, i know AMD are a bit behind in IPC and clocks but i didn't think it was that much..
That would mean Zen 2 would be catching up to Haswell.
That would mean Zen 2 would be catching up to Haswell.
Associate
- Joined
- 21 Sep 2018
- Posts
- 895
The Aida Mem Test, like one comment said, was way off. More like the RAM was set at 2400MHz CL16.
The rest to the tests somewhat reflect the RAM performance.
3200 CL16 DOCP with Gen 1.
https://i.imgur.com/lsoP9ho.jpg
The rest to the tests somewhat reflect the RAM performance.
3200 CL16 DOCP with Gen 1.
https://i.imgur.com/lsoP9ho.jpg
Last edited:
Associate
- Joined
- 9 Jan 2019
- Posts
- 885
Zen is no SB level ipc and if you think that you show how little you know about cpu tech.
Said while quoting an equally "biased" outlier benchmark. I dunno what this guy's methodology is but other IPC tests do not have the same conclusions. Maybe he's using slow RAM, I dunno.
Caporegime
Something wrong with that...
As i wrote on Reddit.
------------
I was surprised to see TheGoodOldGamer latests IPC video in which his results show Zen+ IPC level some 15% lower than Haswell, he thinks its even lower for Coffeelake.
I cannot get my head around how he got these results given that everyone else who i have seen do the same thing with the same applications result close to Coffeelake, it just blows my mind how he can get results so wildly different to everyone else.
His Cinebench R15 scores are about 15% behind Haswell, his CPU-Z results about the same level of difference.
His Cinebench R15 results he claims with all CPU's at 3Ghz he thinks 122 for Haswell at 3Ghz and 112 for Zen+ https://i.imgur.com/ygezslY.png
Guru3d results:
4790K boosts to 4.2Ghz single core, they result 173 at that, 40% higher and that matches the clocks difference, 40% higher score vs 40% higher clocks.
But at 4.4Ghz the 2700X scores 183, 47% higher clocks vs 3Ghz but the 2700X is scoring 63% higher than his result with clocks normalised, so there is about 15% scores missing from his results on Zen+.
A far easier way is TechSpot, they did them all at 4Ghz.
See that ^^^ with everything at 4Ghz Zen+ is about 97% of Coffeelake, that compliments the Guru3D result....
I ran Cinebench R20 with a Coffeelake owner on another forum, and again his Coffeealke clock for clock was only about 3% faster.
https://forums.overclockers.co.uk/t...enchmark-thread.18849380/page-3#post-32566035
So how the hell is Chris getting these results on Zen+?, they are about 15 to 20% lower than reality.
CPU-Z exactly the same, his results are about 20% lower than they should be....
As i wrote on Reddit.
------------
I was surprised to see TheGoodOldGamer latests IPC video in which his results show Zen+ IPC level some 15% lower than Haswell, he thinks its even lower for Coffeelake.
I cannot get my head around how he got these results given that everyone else who i have seen do the same thing with the same applications result close to Coffeelake, it just blows my mind how he can get results so wildly different to everyone else.
His Cinebench R15 scores are about 15% behind Haswell, his CPU-Z results about the same level of difference.
His Cinebench R15 results he claims with all CPU's at 3Ghz he thinks 122 for Haswell at 3Ghz and 112 for Zen+ https://i.imgur.com/ygezslY.png
Guru3d results:
4790K boosts to 4.2Ghz single core, they result 173 at that, 40% higher and that matches the clocks difference, 40% higher score vs 40% higher clocks.
But at 4.4Ghz the 2700X scores 183, 47% higher clocks vs 3Ghz but the 2700X is scoring 63% higher than his result with clocks normalised, so there is about 15% scores missing from his results on Zen+.
A far easier way is TechSpot, they did them all at 4Ghz.
See that ^^^ with everything at 4Ghz Zen+ is about 97% of Coffeelake, that compliments the Guru3D result....
I ran Cinebench R20 with a Coffeelake owner on another forum, and again his Coffeealke clock for clock was only about 3% faster.
https://forums.overclockers.co.uk/t...enchmark-thread.18849380/page-3#post-32566035
So how the hell is Chris getting these results on Zen+?, they are about 15 to 20% lower than reality.
CPU-Z exactly the same, his results are about 20% lower than they should be....
Last edited:
Caporegime
Something wrong with that...
As i wrote on Reddit.
------------
I was surprised to see TheGoodOldGamer latests IPC video in which his results show Zen+ IPC level some 15% lower than Haswell, he thinks its even lower for Coffeelake.
I cannot get my head around how he got these results given that everyone else who i have seen do the same thing with the same applications result close to Coffeelake, it just blows my mind how he can get results so wildly different to everyone else.
His Cinebench R15 scores are about 15% behind Haswell, his CPU-Z results about the same level of difference.
His Cinebench R15 results he claims with all CPU's at 3Ghz he thinks 122 for Haswell at 3Ghz and 112 for Zen+ https://i.imgur.com/ygezslY.png
Guru3d results:
4790K boosts to 4.2Ghz single core, they result 173 at that, 40% higher and that matches the clocks difference, 40% higher score vs 40% higher clocks.
But at 4.4Ghz the 2700X scores 183, 47% higher clocks vs 3Ghz but the 2700X is scoring 63% higher than his result with clocks normalised, so there is about 15% scores missing from his results on Zen+.
A far easier way is TechSpot, they did them all at 4Ghz.
See that ^^^ with everything at 4Ghz Zen+ is about 97% of Coffeelake, that compliments the Guru3D result....
I ran Cinebench R20 with a Coffeelake owner on another forum, and again his Coffeealke clock for clock was only about 3% faster.
https://forums.overclockers.co.uk/t...enchmark-thread.18849380/page-3#post-32566035
So how the hell is Chris getting these results on Zen+?, they are about 15 to 20% lower than reality.
CPU-Z exactly the same, his results are about 20% lower than they should be....
CPU-Z
Ryzen 1600 @ 3.85Ghz 440 ST
4790K @ 4.2Ghz (Actual ST Boost) 451.
So with 30% higher clocks i'm scoring 65% higher... yeah, doesn't add up.
Last edited:
Caporegime
I did this one at 3.0Ghz, the same speed as Chris.
343
vs
265
A difference of 30%, anyone with CPU-Z, Ryzen and Cinebench for that matter can easily replicate this.... Chis is 30% down on what he should be scoring at least in CPU-Z, about 20% down in Cinbench, its utter nonsense, i don't know why.
343
vs
265
A difference of 30%, anyone with CPU-Z, Ryzen and Cinebench for that matter can easily replicate this.... Chis is 30% down on what he should be scoring at least in CPU-Z, about 20% down in Cinbench, its utter nonsense, i don't know why.
He's using really slack ram timings, might be it.
Caporegime
No ^^^ he gives his RAM as 3000Mhz CL16, the same as mine.
--------------------
Cinebench R15
Zen / Haswell @ 3.0Ghz
Chris 108
Me 122 +13%
I think Chris's Haswell scores are higher than they shoul'd be at 3.0 Ghz, Hasswel at the top there scoring 165 at 4.4Ghz, my own Haswell score there 162 at 4.5Ghz.
He says Haswell scores 122 at 3.0Ghz, the clock diffrence between 3.0Ghz and 4.4Ghz is 47% yet the score diffrence between his claimed 3.0Ghz and actual 4.4Ghz on the same CPU is only 35%, a 12% discrepency.
His results for all the Zen scores are total BS, they are much lower than they should be and his Haswell scores are 12% higher in Cinebench than they should be.
--------------------
Cinebench R15
Zen / Haswell @ 3.0Ghz
Chris 108
Me 122 +13%
I think Chris's Haswell scores are higher than they shoul'd be at 3.0 Ghz, Hasswel at the top there scoring 165 at 4.4Ghz, my own Haswell score there 162 at 4.5Ghz.
He says Haswell scores 122 at 3.0Ghz, the clock diffrence between 3.0Ghz and 4.4Ghz is 47% yet the score diffrence between his claimed 3.0Ghz and actual 4.4Ghz on the same CPU is only 35%, a 12% discrepency.
His results for all the Zen scores are total BS, they are much lower than they should be and his Haswell scores are 12% higher in Cinebench than they should be.
Soldato
3 posts in a row........
Someone is defending hard
Someone is defending hard
Soldato
3 posts in a row........
Someone is defending hard
Very little to defend, the guy who made the video doesn't seem to have any clue what he is doing.
*Posted from my i7 4770.
Caporegime
3 posts in a row........
Someone is defending hard
At least he’s posting some evidence. I thought Zen was above Sandy IPC from previous reviews.
Caporegime
3 posts in a row........
Someone is defending hard
But its so easy to debunk....
Associate
- Joined
- 9 Jan 2019
- Posts
- 885
He has been fiddling with the memory timings and god knows what, on a cpu architecture like Zen where timings and memory are critical to IF its not a good idea.
So yeah you can get Zen to run as slow as Haswell or SB, but you need to cripple it to do so. Meanwhile with intels stuff you just need to install the latest security patches to horse its performance
So yeah you can get Zen to run as slow as Haswell or SB, but you need to cripple it to do so. Meanwhile with intels stuff you just need to install the latest security patches to horse its performance
Caporegime
Both are true.
Caporegime