-
Posts
634 -
Joined
-
Last visited
-
Days Won
10
Content Type
Profiles
Forums
Events
Blogs
Posts posted by ground
-
-
@antome has a search thread maybe this fits
-
1
-
-
Anyone considered Threadripper for 32M? It doesn't seem to have as bad fabric coldbug issues as zen2.
-
1
-
-
-
14 minutes ago, gtxx58 said:
Not expected less from you, great stuff, you do something special to run full at full pot ?
Only thing it needed was less PLL (was running 1.4V PLL). Was quite surprised, since at default it coldbugs at -120ish. Or this board just works better on cold then my R2E did, not quite sure yet, but I'm certainly buying a R3F now (this one was a loaner from Gigelz).
Also had a 14.99 score but didn't take a screenshot because I didn't notice my 14.96 was invalid
One of many scores from the session with Bullshooter, Gunzi and Tantekaethe.
-
Agreed, just 2 weeks change could make it a whole lot more convenient. That or 2 weeks longer and a mid/late december background change to make sandbagging harder
-
3
-
-
jup, BCLK didn't scale at all beyond -110 or so
Core is nowhere near the end of scaling but BCLK just won't do more.
-
400 with bsel and the rest with setfsb I guess?
-
44 minutes ago, MrGenius said:
What's the deal with the number of scores for 3DMark06 LEGACY GPU? It says "TOP 5 AVERAGE". But no country has any more than 4 scores yet. And when I subbed my score for United States, which already had 4 scores, it kicked @jpmboy's score off the list. Leaving 4 scores still. So how does a country ever get 5 scores on the list? And...NO...his score was with a GTX 1070 Ti, my score is with an RX Vega 64. So that's not it. This makes no sense. Am I missing something?
PCI, AGP, IGP, Hybrid and PCIe. Germany is missing AGP, Italy PCI for example.
-
1
-
-
Who needs skulltrail?
-
what voltage was this run at?
-
I knew someone would take on this score eventually, great work! Will be harder now...
-
1
-
-
wow, 700+ FSB on an 98 board? apparently not all boards later then 8B are bad...
-
24 minutes ago, Mythical tech said:
I think what we need is an extremely budget dice pot even if it is not amazing having a low cost entry is a way to get more people to stick with it. I made my first dice pot a few years ago out of the copper slug from an intel stock cooler and a copper pipe that i joined jbweld. If there was a pot on the market for sub 150$ I think a lot of people would find it more realistic to invest in the hobby without spending 350$ on a new high end pot.
yeah used pots can be found for <100€ at times but they are usually sold fairly quickly and new ones are really expensive. Still using an old Ryba pot I picked up last year for 80€. Did spend significantly more on Dice and LN2 since then though...
-
pm for #80 and #84
-
14 minutes ago, unityofsaints said:
That's not the PCI version. Or is GPU-Z detecting the interface wrong?
It's detecting the interface wrong. Happens with HD 5450 DDR2 as well. Just look at a random score for any of the other semi-modern PCI GPUs, they all show up in the same way. Try finding a score where its showing up differently
-
-
Just now, Strunkenbold said:
I didn't read through all of this, so this may have already answered. Stage 10 is divided by CPU socket. So Mobile chips can be used. There are some fancy Chinese CPU creations, having a mobile chip on a shim for LGA socket. It's basically a form of BGA to LGA adapter. Running a mobile chip in a desktop motherboard can yield some slight advantage, especially if you plan to run them under cold.
But technically those are custom creations and thus are actually falling under the retail available hardware rule. On the other hand, the CPUs itselfs where retail available. It's just the shim that was created by third party and they are available to anyone who can organize the shipping from China. I just want to double-check that, hence I ask.
as cool as they are, the few I've seen in the wild so far were usually ES chips - thus not viable for comps. IF you find these with retail chips the question applies off course.
-
Isn’t a team named after a large youtube channel somewhat complicated if its not represented by at least one working for them?
-
1
-
-
4 hours ago, Luebke said:
@mickulty: the problem is not, that i would prefer legacy pci-e hardware, the problem is, modern GPU´s make way more points so the results of legacy GPU nearly does make no difference. the relation is disproportunal. if our team had a very good 2080Ti i could run my 3850 easily @ stock to beat macsbeach98´s team (btw very nice work macsbeach98). that´s a pity to me...
just my two cents worth...
Lets have a look at the last time there was a "GPU socket" categorie and look at #1 and #2...
https://hwbot.org/search/submissions/permalink?ids=2467844,2468072,2468323,2462342,2468134
https://hwbot.org/search/submissions/permalink?ids=2468318,2468259,2448048,2450594,2468354
And, for better comparision, the full results.
https://hwbot.org/competition/country_cup_2013/stage/1157_3dmark03_(legacy)
Looking at this even the PCI score could end up relevant this year... Back then if Germany would've had a higher score in PCI they would've beaten Australia in that stage... So shut up with "the old stuff is irrelevant". If we only focus on the 2080 Ti score, there is no way we are winning that stage even if we have fillers...
-
2
-
3
-
-
3 minutes ago, ObscureParadox said:
Why does the cascade have a switch with two "on" options
Maybe its a no-no switch?
-
sweet CPU pot tagg
-
:D It seems to have worked!
-
1
-
-
Nice chip :D
-
1
-
-
Some boards just seem to hate superpi, when I was testing my P5E3 Premium last month with NB on SS I could validate 670+ easy, run 1M at 660 with decent memory but couldn't manage 32M above ~615 no matter what I tried. Memory tested at 700ish, 900ish and >1030 without issues, I kinda think its just me...
OnisA - X7DAE @ 517.1MHz - 517.1 MHz Reference Frequency
in Result Discussions
Posted
These 771 boards are quite fun, though I feel like either CPU-Cool works significantly better then setfsb on these boards or I was doing something else wrong.
I would for sure like to further explore these supermicro boards next year, if you want we can make a discussion thread about these boards at some point. Even going cold with memory known to do 500 MHz in another board I couldn't get close with NB on LN2 (actually barely got anything more over ambient anyways), so great work!