Printable View
@chew
Can you change tRC in bios
Because the PII Need less CPU-NB whan tRC was highest and so they were much cooler and get higher clocks
Also prefere the CPU-NB from less RAM voltage and need les CPU-NB whan it was lowest
And the FX needs less V-Core if tRC was highest
Also a Little bit less CPU-NB, but not so much like the PII
Upon investigation i can find nothing wrong with psu and hard drive.
I also see no signs of any failed components on the board.
Think it was just a freak accident. Some chip was probably partially defective.
It can still be posted and get into bios by holding power button down so its not all dead. Just half dead.
No clue.
Ill send it back to gigabyte for inspection.
Trc is only changeable by using higher ram divider. Lower divider lower trc higher divider higher trc.
Like most have said on all boards 5 main timings available only.
Maybe You Need less v-Core / CPU-NB with higher dividier
You can test it If you are motivated
Did the bios use than tRC from SPD?
Because Maybe modify the SPD can helps
Nope trc is not matching SPD.
The agesa is controlling it all and it is doing it via memory divider chosen.
Cant test anything atm. Not till i go grab a board but hesitant to do so. Only boards that i can get in stock today that support 3200 memory do not list my sets as qvl.
The others claim 2933 and lower.
OK so that was fun.
Took a trip to microcenter.
http://i1253.photobucket.com/albums/...psem272frv.jpg
Remembered to take a pic of rig on display this time
http://i1253.photobucket.com/albums/...ps4ukqg9wv.jpg
Then I bought this for a few reasons.
#1 dirt cheap there was a MSI tomahawk also but $7 more and then the asrock fatality was like $109 new only
#2 based on qvl I felt it was the best fit for my Samsung based IC's, both asrock and the msi seem to be Hynix biased based on qvl.
#3 I'm grass roots I don't care what it is I will overclock it.
http://i1253.photobucket.com/albums/...pscrnrkvkx.jpg
The NVidia card is for some high PCI endeavors. I love my clearance items.......cheap paste for a lot of cpu swaps. .
The cheaper the better when swapping cpu's out to test stuff.
Same hardware that system died on intentionally to see if something broke and or something caused the other board to break.
I think just the board had a freak accident as all is well.
http://i1253.photobucket.com/albums/...pssq8glzkz.jpg
Up and running in no time.
http://i1253.photobucket.com/albums/...pssabfiqlp.jpg
Now sometimes its not the size that matters but the functionality and bios tuning.
I'm not telling you all to go out and buy this and you can run 4.1
All I'm saying is I think the top tier boards with lackluster memory support better get there acts together.
http://i1253.photobucket.com/albums/...psvli7vao4.jpg
http://i1253.photobucket.com/albums/...psc60p0mp1.jpg
:eek:
What were they cooling the CPU with in the Microcenter build? That is one of the ones I am competing against lol
ek block on water
Wow, mem scores are impressive. I can't catch that with Hero even with better clocks and tighter timings.
Can you run MemTweakIt and show the subtimings?
It seems there's much to be optimized on the Crosshair board.
This is by far the best performance I've seen so far.
PS: Actually I forgot to set Performance Bias to aida/geekbench in bios. Now it is much better.
Edit 2: hmm, I think it ight be a bug - got it once to 67GB read, once to 99GB, but other times it is similar to what I've tested before, no matter performance bias option in bios.
Are the 1.069v real?
No its about 1.475 the voltage is reporting wrong. They know so should be fixxed soon.
Will run it back to back double check the results. I set my timings manually. Docp was causing me cold boot problems with my oc settings. I think the auto docp paired with manual cpu oc was having conflicts. I set manually boots every time now.
No debug and my port 80s are not working makes for fun diagnotics.
Currently as it stands based on research of boards, manuals, specs, performance, bios i stick by my reccomendation of using a 1700.
What i would change suggestion wise is if running stock speeds and not planning to overclock with a single gpu....screw x370 boards.
Go buy a b350 based on the features that fits your needs best. I think you could even get away with a reasonable oc maybe 3.7-3.8. But 4gig 24/7 well just look at pwm...your asking for problems..
Total cost of this particular setup is $550 cpu motherboard and memory....i think the fury X costs more than the platform by itself..
Might these B350 be OK to hold a 6C/12T at ~3.9-4.0 (assuming they'll overclock the same, not more) during prolonged stress testing?
Honestly I'm a bit surprised that there are no B350 boards with more/redundant phases, unlike what happened before on entry 990FX/X or top 970 models at a <130e budget. These had 8+2 or 6+2 pwm setups inherited from previous 890FX models, while the newer pricier top FX models had revised pwm to account for bulldozers. It's weird to see only ~4+2 AM4 boards on that price range now.
Tough call. I still probably would not attempt prime 95. What you could do is run cinebench back to back to back. when you don't crash anymore drop it 100 mhz. call it stable. 3 chips have all done prime 95 100mhz under my solid back to back cinebench spam.Quote:
Originally Posted by AlleyViper
Ok infrared
Heres your tests.
First 3 back to back, the 2nd and 3rd was done with cpu-z and memtweakit and maxmem open.
http://i1253.photobucket.com/albums/...pscisxr2nb.jpg
http://i1253.photobucket.com/albums/...psz4fmzwqr.jpg
http://i1253.photobucket.com/albums/...psf1etmavg.jpg
This was done next after reboot just trying to find a inconsistency between boots or whatever way possible.
http://i1253.photobucket.com/albums/...ps1heghojj.jpg
Then I tried sucking in timings.......seems at 32x cas latency is a victim of an autorule.......I have seen this on more than 1 board.....so most likely (AGESA) can not do 12 or 10 or 11. 13 works but does not apply.
Seems the platform likes even timings vs odd......still up for debate as I test across boards.
http://i1253.photobucket.com/albums/...ps20bcrvfz.jpg
http://i1253.photobucket.com/albums/...pszxs2qi0w.jpg
Thanks.
Unfortunately my board (Crosshair VI) seems to be dead. Lights up the buttons, but does not react on anything. Bios flashback button doesn't work too (no light).
Was that under ln2 or pre ln2?Quote:
Originally Posted by I.nfraR.ed
LN2, but I was just starting to see what the chip likes.
Nothing was wet, not visually burnt elements, voltages were not that high - 1.6V cpu, 1.2V SOC, 2.2 PLL, 1.3V SB.
I updated to 0902, which is supposed to fix the brick situation. After a failed overclock I rebooted manually, then it shut off and now nothing happens when I press the Power button.
Teared the rig down, everything was nice and dry. Bios flashback button does nothing, does not even lights up. Removed every component - cpu, ram, ssd, vga, usb devices, so it is just a bare board with PSU connected - no reaction. PSU is fine, just checked with Apex. Hopefully it is just the board, because the CPU seems promising.
Not sure if that's the EC firmware corrupt or something else.
Well as ive been saying there are like two bios one is on chip one is vendor. As you beat the cpu up on chip who knows what can happen.
I will obviously get to this eventually but for now i think i will wait and help iron out the problems for the mainstream.
The more reliable it is at normal temps hopefully will carry over to cold.
FUGGER writes about something similar on his board - no post when hitting power button and it did similar thing to mine before dying.
I just really hope the cpu is fine, will return the board for RMA obviously, but there's no stock at the moment, so don't know how much I will have to wait :(
Weird stuff happens with these cold apparently.
Can anyone do a prime95 1 hour run and report back with temps please?
Small FFT
Thanks
http://i.imgur.com/FAyUq9i.gif
On a b350?Quote:
Originally Posted by easyrider
the question is rather vague volts speed etc......
on x370 ive done hours w/temps in pics.......just look through thread.