Ryzen - Return of the Jedi

[chew*]

You under cascade fugger- 80ish? Read my ryzen info part of thread.

[chew*]

My normal 24/7 cooling setup^^
Works well since 939
Never distroy an MB, RAM ore CPU

That would be a smart common sense and realistic way to use the g5 in a case.

Keep in mind i test on a bench and i always must take into account that the community will be in a case.

If it hot on the bench a pc case real world is hotter.

[Hell Hound]

can you vsr to 4k w/ the 290x's and note min fps please,I haven't seen multi monitor/gpu test

[techtrancer]

That would be a smart common sense and realistic way to use the g5 in a case.

Keep in mind i test on a bench and i always must take into account that the community will be in a case.

If it hot on the bench a pc case real world is hotter.

Think most of the new MB has this Problems
Bad at default and dangerous with OC
MB in the far away past often has installed Fans on them
But without they are Maybe 0,01$ cheaper in producing ^^

For that are on the new more led the nobody is dived. And more spreader that heat accumulates more than cooling but looks good ^^

[chew*]

Trying to pull a blanket over any vendors short comings in an effort to get continued free sample support does not help community nor does it spark innovation or changes.

I once told community publically during p67 craze that one board was only good for the motherboard box it came in to bench on top of. It died under rather light use.

Brutal honesty = innovation

Blankets = stagnation.

[techtrancer]

Think the DFI UT 790FX-M2R has One Of the best Voltage converter cooling for case use ever
And at 939 Time, most of the MB has Fan in the NB
And now everything is fanless
In my opinion not Innovation, not stagnation but backward

[chew*]

Well they read forums more than most think. They are probably already in design on the next version gen whatever.

The time to be heard is now. They wont just change stuff once its laid out. To costly.

[Raiderman]

Hope you dont mind Brian, but I have been quoting some of your findings at notebookreview.com Of course you will have to wade through the earlier posts.

http://forum.notebookreview.com/thre...ly-zen.799348/

[techtrancer]

Well they read forums more than most think. They are probably already in design on the next version gen whatever.

The time to be heard is now. They wont just change stuff once its laid out. To costly.

We well see in the Future, if we get good cooled MB with Fans again
I also can live without onboard Fans
With wather cooling it's easy to install them selve
But mostly use Air cooling, and mostly of them has Problems with to Hot NB & voltage converters and finaly unstable systems and dying MB

[chew*]

Np Raiderman.

Welcome to port questions here ill answer them.

You can inform the user who said giga does 3400 with latest bios as innaccurate at least for g5. I have the latest beta possible from hq.

32x max no ref clock nor can ref clock be unlocked...it lacks physical hardware chip to do so.

You can inform the other person who said that smt disabled bug no issue can use ryzen master to change multi on core 0 as innacurate also. Yes hpet was enabled for ryzen master.

Might want to let them know my skill/knowledge lvl at ocing, effeciency, analysis and in general after reading some posts. Chuckled a few times at some comments.

I need to update that info upon verifying it is not a workaround for 1700 although it should have worked...

[Ket]

Trying to pull a blanket over any vendors short comings in an effort to get continued free sample support does not help community nor does it spark innovation or changes.

I once told community publically during p67 craze that one board was only good for the motherboard box it came in to bench on top of. It died under rather light use.

Brutal honesty = innovation

Blankets = stagnation.

Agreed! A fair few manufacturers dislike me for taking this stance, particularly Sapphire, OCZ and Asus Given the demolition OCZ went under a while back though I'd be surprised if anyone is left there now from when that bad blood kicked up. Karmas a ain't it, OCZ?

[techjesse]

Great Review chew Now Ln2 Benches

[techtrancer]

@chew

Have you run any Benchmark under w7 also?
I read that w7 ist with Ryzen in different Benchmark up to 13% faster than w10 at the Moment
The stilt Write that in w10 is any Plugin who opimice w10 for the FX
So Ryzen Runs at the Moment with the FX Plugin and Plugin for Ryzen should come in around 1 month from MS

[chew*]

Have not yet but i can. Running the timespy comparison i referenced in review. Waiting on bios to verify some stuff. Probably go grab some hd's this weekend for multi OS testing.

Time spy updated. Post #12

http://www.xtremesystems.org/forums/...=1#post5254468

[techtrancer]

am i right that Ryzen scales good with low Temperature and Bad with v-Core?

[chew*]

Looks like Ryzen scales good with low Temperature and Bad with v-Core?

ambients were 20C lower needed a little more cpu power for the comparison.

def not sub 0 idle was 40c. voltage is finicky like I said more crashes faster.

Fury x was cold though, dunked radiator in ice water.......didn't know they had cold slow bug, that made things a tad tricky lol.

Did not feel like flashing bios to card.

[techtrancer]

did you plan also to test ryzen without HS?

[chew*]

probably not. Its never benefitted cold in the past and I have a heavy load ln2 pot that I think can tame it but would certainly crush core.

[SIOUX]

Lots of good info Chew..thx a lot

[chew*]

Updated mother board rants. Tagged update in bold for easy reference.

http://www.xtremesystems.org/forums/...=1#post5254424

For a blast from the past here is the main link to the gallery but the gallery is huge and kinda hard to navigate.

http://www.chew.ln2cooling.com/

[chew*]

@ raiderman

For you buddies on notebook forum

Key thing to note is the best result was achieved by running bclock up with 2400 ram divider.

NB still is 1/1 with ram this way. Doing this tricks the timings that we have no access to into running tighter.

I suspected as much but since the board i'm stuck on for the moment lack bclock / ref clock whichever you prefer I could not confirm it. It's obviously confirmed now.

That is part of the performance deffeciency.

Furthermore yes there appears to be a scheduling issue in 10.....I personally have not upgraded myself to 10 ( I'm not a fan of anything MS till it has a SP1 under its belt ) but I have a copy for review purposes.

Those testing Win 7 are reporting 7-15% increase in performance.

I'm picking up a stack of spare hd's tomorrow to run multiple operating systems and compare MY own intel system in 3d applications across multiple OS.

Its antiquated by todays standards ( p67 2600K ) but it can do 4.5 4c/8t and 5.0 4c 24/7 reliably.

Pictures courtesy of elmor.

[chew*]

can you vsr to 4k w/ the 290x's and note min fps please,I haven't seen multi monitor/gpu test

Totally missed this. I can get to the x fire 290xs eventually and i do have the right equipment to prevent throttling but its gonna be awhile. Must dismantle the case and water loop. Some idiot shoved it all in a 450D and made it fit crappy pic

Mine are not oced per say. I use LCS bios 1060/1350 and leave them alone. Both are high asic and do this without voltage. Peak is 1120/1500 no voltage which is why i use that bios. Puts me dead center middle of max stable.

Cant remember if my acer KG240A works @ 4k vsr but i will check.

You should keep in mind though that min gets worse oced max goes up when you get around throttling oced.

Tomb raider (rotr) benchmark is notorious for this. Drove me nuts

[techtrancer]

Quote of TheStilt

In 3D testing I did recently on Windows 10, the title which illustrated the biggest SMT regression was Total War: Warhammer.

All of these were recorded at 3.5GHz, 2133MHz MEMCLK with R9 Nano:

Windows 10 - 1080 Ultra DX11:

8C/16T - 49.39fps (Min), 72.36fps (Avg)
8C/8T - 57.16fps (Min), 72.46fps (Avg)

Windows 7 - 1080 Ultra DX11:

8C/16T - 62.33fps (Min), 78.18fps (Avg)
8C/8T - 62.00fps (Min), 73.22fps (Avg)

[FlanK3r]

Briand, can you try last beta of AIDA, if memory tests are now fixed? (my CPU coming tomorow, but tomorow I must also finish some work with 7700K and Apex board)

http://www.xtremesystems.org/forums/...37#post5254837

[chew*]

Would not surprise me but vulkan api/dx12 has everyone rushing to 10. Least on the amd front. Nvidia guys are not liking dx12 to much.

Yes flanker ill check it out. but most testing will be on hold till i get back from a trip. Have to shoot to PA overnight.