AMD,socket AM3 Motherboard

[bullant]

Hey guys,just looking to get a AM3+ board and was wonder what would be a good choice,I know the ASUS Crosshair V Formula-Z Motherboard is a nice board but was wondering how does the ASRock 990FX Extreme9 Motherboard compare to the Asus board.Will be mainly using the board for extreme 2D benching,any help be great

Thanks bullant

[Evantaur]

do you mean AM3+?

[bullant]

Yes sorry AM3+

[AMDforME]

The Asrock E9 is a slightly updated version of the Asrock 990FX Fatality mobo. Both work quite well and are a far better value than the over-priced Asus. In the Fatality thread below D1nky does a lot of competitive benchmarking so you might want to chat prior to purchase.

http://www.overclock.net/t/1078815/p...ofessional/390

FWIW, the Asrock E9 and 990FX Fatality are some of the very few mobos capable of running the FX-9000 series without issue as their 12-phase VRM circuit is very capable unlike most of the other AM3+ mobos which are limited to ~160w max. As far as CPU OC'ing the Fatality and E9 also seem to be as good if not better than the Asus mobos. As with all OC'ing, YMMV.

[Ace123]

I have an ASUS M5a99X Evo and it makes this puny fx-4100 do 5.1ghz with DDR3 at 2550mhz..... seems to be a good board.

[bullant]

The Asrock E9 is a slightly updated version of the Asrock 990FX Fatality mobo. Both work quite well and are a far better value than the over-priced Asus. In the Fatality thread below D1nky does a lot of competitive benchmarking so you might want to chat prior to purchase.

http://www.overclock.net/t/1078815/p...ofessional/390

FWIW, the Asrock E9 and 990FX Fatality are some of the very few mobos capable of running the FX-9000 series without issue as their 12-phase VRM circuit is very capable unlike most of the other AM3+ mobos which are limited to ~160w max. As far as CPU OC'ing the Fatality and E9 also seem to be as good if not better than the Asus mobos. As with all OC'ing, YMMV.

Thanks for reply,some very useful information,I was thinking of getting the FX 9370

I have an ASUS M5a99X Evo and it makes this puny fx-4100 do 5.1ghz with DDR3 at 2550mhz..... seems to be a good board.

Thanks for the info

[Darxide]

I have the Crosshair V Formula Z, it's an awesome board, but pricey.
I'll be honest, my reasons for choosing it where entirely cosmetic. I wanted the whole black/red theme on this build.

[I.nfraR.ed]

You should look at the high-end boards only.
Although last Asus board I bought (mainly because there are Gigabyte competitions on the bot) was Crosshair IV Formula, I think CVF-Z is the best choice.
If you want to go cheap, maybe get 990FXA-UD3, but I hate all these revisions of the same board Gigabyte releases with bioses not compatible across different versions.
I don't even expect future bios support for older revisions. UD3 has 4 different revisions! It would be fair if Gigabyte offers us to return the old board and get a new one, but ofcourse that won't happen.
If something is not right with the board and they need a new hardware revision then where's my fault?

I'm sick of this, otherwise I would recommend you a GB board.

Asrock - don't have any personal experience with their newer boards, but might be good. Asrock progressed a lot in the last 2 years, but I think Asus is still the leader in AMD.

[Barr3l Rid3r]

You should look at the high-end boards only.
Although last Asus board I bought (mainly because there are Gigabyte competitions on the bot) was Crosshair IV Formula, I think CVF-Z is the best choice.
If you want to go cheap, maybe get 990FXA-UD3, but I hate all these revisions of the same board Gigabyte releases with bioses not compatible across different versions.
I don't even expect future bios support for older revisions. UD3 has 4 different revisions! It would be fair if Gigabyte offers us to return the old board and get a new one, but ofcourse that won't happen.
If something is not right with the board and they need a new hardware revision then where's my fault?

I'm sick of this, otherwise I would recommend you a GB board.

Asrock - don't have any personal experience with their newer boards, but might be good. Asrock progressed a lot in the last 2 years, but I think Asus is still the leader in AMD.

I guess there is nothing wrong with them... newer revisions like 3.1 has throttle problems. I still having my UD3 Rev. 1.0 for almost 3 years now and it still rocking, sam OC abilities.

[Daveburt714]

Hi Bullant,

I have both boards, although my Crosshair V is not the Z variant (just the regular formula).
As far as OC'ing goes I prefer the Asus board hands down! IMHO, the price premium is worth it for overclocking.
It also has the added benefit of having RCTweakit which allows you to clock the rig from a separate machine via a USB cable.
This comes in very handy when trying to dial the rig in, or to help squeeze the last bit out of it...

The ASRock EX9 is also a nice board, has a lot of nice features and a very strong power section. My biggest problem with it (still) is the bios.
The best word I can think to describe it is "fidgety"... The original bios version (1.0) was really bad and wasn't updated for over 6 months.
The newest release (1.40) solved a lot of problems, but still has that same feel. If you decide to go the EX9 route make sure your on the newest bios from the start!
You can get good clocks out of it, you just have to work harder and fiddle with more settings to get there...

Which ever way you decide to go, Have fun and happy OC'ing!

BTW, The Crosshair also supports the 9xxx series chips:
http://www.asus.com/ROG_ROG/CROSSHAI...Z/#support_CPU

[BeepBeep2]

You should look at the high-end boards only.
Although last Asus board I bought (mainly because there are Gigabyte competitions on the bot) was Crosshair IV Formula, I think CVF-Z is the best choice.
If you want to go cheap, maybe get 990FXA-UD3, but I hate all these revisions of the same board Gigabyte releases with bioses not compatible across different versions.
I don't even expect future bios support for older revisions. UD3 has 4 different revisions! It would be fair if Gigabyte offers us to return the old board and get a new one, but ofcourse that won't happen.
If something is not right with the board and they need a new hardware revision then where's my fault?

I'm sick of this, otherwise I would recommend you a GB board.

Asrock - don't have any personal experience with their newer boards, but might be good. Asrock progressed a lot in the last 2 years, but I think Asus is still the leader in AMD.

+1 to this.
If trying to overclock for extreme 2D benching then I would not buy a Gigabyte board at all though, except maybe the UD7. I personally hate how Gigabyte seems to make users beta testers time and time again with their boards. I'm not too happy with my UD5 either, and Gigabyte's overclocking software (EasyTune) is atrocious.

My Crosshair V Formula has been a really strong board, and RC TweakIt makes overclocking much more streamlined, plus more in-depth options for tweaking PWM settings, memory settings, etc, and overall (IMHO) is much easier to deal with after a hard reset or incompatible settings. It will reset itself the way you want every time, when settings are unstable, keeping settings in the BIOS, when the Gigabyte board loops or sits forever until you clear CMOS or will reset every setting in the BIOS when it comes back.

Haven't tried ASRock yet outside my M-ITX board (which I am unhappy with), but since they picked up NickShih they have made a lot of improvements, but hard to say if they are as good as ASUS yet.

[bullant]

I have the Crosshair V Formula Z, it's an awesome board, but pricey.
I'll be honest, my reasons for choosing it where entirely cosmetic. I wanted the whole black/red theme on this build.

Yeah black and red is a nice colour

You should look at the high-end boards only.
Although last Asus board I bought (mainly because there are Gigabyte competitions on the bot) was Crosshair IV Formula, I think CVF-Z is the best choice.
If you want to go cheap, maybe get 990FXA-UD3, but I hate all these revisions of the same board Gigabyte releases with bioses not compatible across different versions.
I don't even expect future bios support for older revisions. UD3 has 4 different revisions! It would be fair if Gigabyte offers us to return the old board and get a new one, but ofcourse that won't happen.
If something is not right with the board and they need a new hardware revision then where's my fault?

I'm sick of this, otherwise I would recommend you a GB board.

Asrock - don't have any personal experience with their newer boards, but might be good. Asrock progressed a lot in the last 2 years, but I think Asus is still the leader in AMD.

Thanks ,Yep I understand what you are saying,yes very true asrock has come a long way in last 2 years

Hi Bullant,

I have both boards, although my Crosshair V is not the Z variant (just the regular formula).
As far as OC'ing goes I prefer the Asus board hands down! IMHO, the price premium is worth it for overclocking.
It also has the added benefit of having RCTweakit which allows you to clock the rig from a separate machine via a USB cable.
This comes in very handy when trying to dial the rig in, or to help squeeze the last bit out of it...

The ASRock EX9 is also a nice board, has a lot of nice features and a very strong power section. My biggest problem with it (still) is the bios.
The best word I can think to describe it is "fidgety"... The original bios version (1.0) was really bad and wasn't updated for over 6 months.
The newest release (1.40) solved a lot of problems, but still has that same feel. If you decide to go the EX9 route make sure your on the newest bios from the start!
You can get good clocks out of it, you just have to work harder and fiddle with more settings to get there...

Which ever way you decide to go, Have fun and happy OC'ing!

BTW, The Crosshair also supports the 9xxx series chips:
http://www.asus.com/ROG_ROG/CROSSHAI...Z/#support_CPU

Thanks,yeah I have seen the bios support on the Asus is really good,nice to know about asrock bios

+1 to this.
If trying to overclock for extreme 2D benching then I would not buy a Gigabyte board at all though, except maybe the UD7. I personally hate how Gigabyte seems to make users beta testers time and time again with their boards. I'm not too happy with my UD5 either, and Gigabyte's overclocking software (EasyTune) is atrocious.

My Crosshair V Formula has been a really strong board, and RC TweakIt makes overclocking much more streamlined, plus more in-depth options for tweaking PWM settings, memory settings, etc, and overall (IMHO) is much easier to deal with after a hard reset or incompatible settings. It will reset itself the way you want every time, when settings are unstable, keeping settings in the BIOS, when the Gigabyte board loops or sits forever until you clear CMOS or will reset every setting in the BIOS when it comes back.

Haven't tried ASRock yet outside my M-ITX board (which I am unhappy with), but since they picked up NickShih they have made a lot of improvements, but hard to say if they are as good as ASUS yet.

Thanks,yes I also agree with you on the Gigabyte,so I guess the Asus board is the best option seeing not to many people have tried the Asrock board and Asus does update bios often

Thanks for all the info guys

[AMDforME]

Asrock does frequent BIOS updates whenever there is a need for one. Anyone who's worked with Asrock in recent years can confirm this. Many folks do not yet understand that excessive vcore on the Vishera model CPUs causes instability. This has never been the case in the past unless the CPU was running hot, but it is the case with Vishera CPUs IME and that of other folks who test carefully. Running too much vcore instead of lowering the vcore and upping the CPU NB, and HT voltage, will cause instability and lower OC's.

IME the Gigabyte mobos are very reliable and the dual-BIOS eliminates any BIOS hassles when you OC beyond the range of the hardware. The VRM circuit on all of the Gigabyte AM3+ mobos is insufficient for serious OC'ing of FX CPUs however, other than the just released Giga model designed specifically for the FX-9000 series processors.

Asus mobos have a lot of adjustments which many folks perceive as making them better at OC'ing, but actual test results show these additional adjustments don't necessarily improve OC'ing performance at all. IME Asus tech support is laughable and their warranty support is worse. They are downright insulting and send manually repaired boards as a replacement for a virtually new mobo that goes bad. Asus is using multiple companies to manufacture their mobos so there is no telling what you will get when you purchase retail. The hardware review sites all get hand picked, pre-tested Asus mobos to insure favorable reviews.

[I.nfraR.ed]

I guess there is nothing wrong with them... newer revisions like 3.1 has throttle problems. I still having my UD3 Rev. 1.0 for almost 3 years now and it still rocking, sam OC abilities.

Can't agree with that, sorry. When you're benching with LN2, missing LLC kills my OC. I have to set 1.95V just to get 1.9V under single thread load, not to mention the 0.7V drop in multithread.
I'm risking my chip just to get the desired much lower voltage. All rev1.0 boards had the same problem, even the UD7 which is unacceptable.
In this case, the normal thing would had been a full replacement (like the Intel P67 chipset), it is a technical flaw after all or at least that's how I see it. However I understand no vendor will do it this way.
They fixed LLC in rev3 of UD3. What I like about the non-EFI revisions of the boards is efficiency, but that's all.
And I actually like the 970A-UD3 better, although it can't handle a big multithread load, but that's to be expected for a low-cost board.

They also replaced the plastic pushpins replaced with screws just in the 4th revision. Heatsinks fastened with pushpins move even with the slightest hit.

IME the Gigabyte mobos are very reliable and the dual-BIOS eliminates any BIOS hassles when you OC beyond the range of the hardware. The VRM circuit on all of the Gigabyte AM3+ mobos is insufficient for serious OC'ing of FX CPUs however, other than the just released Giga model designed specifically for the FX-9000 series processors.

No. Dual Bios is a good feature, but not when the board thinks your main bios is corrupted, then tries to load from backup bios, but gets to the Gigabyte text only.
And I had to shorten the bios chip pins to force a non-detectable bios and reflash.
So I don't think it works as intended, at least on the UD3 rev 3. It is still better than a single bios, though.

Not saying they are bad boards. It's just that I hate all these revisions, otherwise I generally like them.

[docwhom]

They are downright insulting and send manually repaired boards as a replacement for a virtually new mobo that goes bad. Asus is using multiple companies to manufacture their mobos so there is no telling what you will get when you purchase retail. The hardware review sites all get hand picked, pre-tested Asus mobos to insure favorable reviews.

YES, Asus have definitely slipped in the ranking when it comes to serious overclocking motherboards today. Unfortunately they would rather play the smoke and mirrors game by plying their trade through carefully placed sleeper boards with paid off websites to help promote their so-called "great high end models".

I suppose it promotes their strength through carefully placed propaganda rather than to produce a real serious product range. Their not alone with that strategy when it comes to fooling the idiot public, who no nothing what they are actually buying.

[AMDforME]

I.nfaR.ed-

I've used a lot of Giga mobos and pushed the limits until the hardware crashed and I never had any issue with the dual BIOS including on the UD3 model. I have had to swap BIOS in other brands of mobos when pushing beyond the hardware limit and driving the BIOS into what I call a "stupid loop". Maybe you just had a corrupt BIOS chip?

[BeepBeep2]

Asus mobos have a lot of adjustments which many folks perceive as making them better at OC'ing, but actual test results show these additional adjustments don't necessarily improve OC'ing performance at all. IME Asus tech support is laughable and their warranty support is worse. They are downright insulting and send manually repaired boards as a replacement for a virtually new mobo that goes bad. Asus is using multiple companies to manufacture their mobos so there is no telling what you will get when you purchase retail. The hardware review sites all get hand picked, pre-tested Asus mobos to insure favorable reviews.

lol!

You don't understand the world of competitive extreme overclocking, at all. 100 points in a benchmark of thousands is still 100 points. 1/50th of a second is 1/50th of a second lost. 0.2% efficiency difference is a big deal. Being able to gain 40 MHz because of higher PWM switching frequency is a big deal...

Most settings available on ROG boards that are not available on other models are useless to consumers, period. However there is a difference between normal use and "extreme 2D benching".

My experience with ASUS RMA has been relatively good, they even repaired my MIVE after it died most likely due to condensation.

IME the Gigabyte mobos are very reliable and the dual-BIOS eliminates any BIOS hassles when you OC beyond the range of the hardware. The VRM circuit on all of the Gigabyte AM3+ mobos is insufficient for serious OC'ing of FX CPUs however, other than the just released Giga model designed specifically for the FX-9000 series processors.

As with I.fraR.ed I have had one board brick itself thinking the main BIOS was corrupt when the machine was unstable writing from the backup BIOS, another time refuse to load the backup BIOS, and the new boards seem to be a PITA when unstable on LN2, they sometimes boot trying to re-write (update) the backup BIOS to the version on the main BIOS. That's not to say it is a bad feature, but I much prefer ASUS' implementation which has allowed users to manually select the BIOS chip and when to update the main BIOS...

As for the VRM/PWM situation, Gigabyte's current setup was most likely sufficient to run 9000-series CPUs under normal situations. There is not a radical difference between 8000 and 9000 series, however it seems to me that the UD5/UD7 DrMOS setups are not quite as capable as the traditional setup on the ASUS Crosshair V Formula boards with only 280 amps of total rated continuous output vs. upwards of 1500 amps on the Crosshair V.

I.nfaR.ed-

I've used a lot of Giga mobos and pushed the limits until the hardware crashed and I never had any issue with the dual BIOS including on the UD3 model. I have had to swap BIOS in other brands of mobos when pushing beyond the hardware limit and driving the BIOS into what I call a "stupid loop". Maybe you just had a corrupt BIOS chip?

Under normal use scenarios users shouldn't have any problem with the DualBIOS feature of Gigabyte's motherboards, however in the past boards have been known to fail to flash the backup bios sitting in that "stupid loop" forever, and more recently with the 990FX series boards attempting to update the backup BIOS when leaving BIOS and resetting with unstable settings.

[Barr3l Rid3r]

Can't agree with that, sorry. When you're benching with LN2, missing LLC kills my OC. I have to set 1.95V just to get 1.9V under single thread load, not to mention the 0.7V drop in multithread.
I'm risking my chip just to get the desired much lower voltage. All rev1.0 boards had the same problem, even the UD7 which is unacceptable.
In this case, the normal thing would had been a full replacement (like the Intel P67 chipset), it is a technical flaw after all or at least that's how I see it. However I understand no vendor will do it this way.
They fixed LLC in rev3 of UD3. What I like about the non-EFI revisions of the boards is efficiency, but that's all.
And I actually like the 970A-UD3 better, although it can't handle a big multithread load, but that's to be expected for a low-cost board.

They also replaced the plastic pushpins replaced with screws just in the 4th revision. Heatsinks fastened with pushpins move even with the slightest hit.



No. Dual Bios is a good feature, but not when the board thinks your main bios is corrupted, then tries to load from backup bios, but gets to the Gigabyte text only.
And I had to shorten the bios chip pins to force a non-detectable bios and reflash.
So I don't think it works as intended, at least on the UD3 rev 3. It is still better than a single bios, though.

Not saying they are bad boards. It's just that I hate all these revisions, otherwise I generally like them.

Sorry Infra you are wrong. Only UD5 and UD7 Ver 1.0 has vdrop problems, actually my UD3 overvolt 0,03V while in load state.

Push pins sux, ok...same for the whole line Heatsinks.

LLC appears at revisions 1.1 for UD5/7 and 1.2 for UD3 not at Rev.3.1. Actually Rev. 3.1 has apm problems that might trhottle back overclocked 8 cores FXs.

Well it's better to have newer revisions than stay with the same option for 3 years or more. You don't see that on intel side cause intel boards are obsolete within a year.

[BeepBeep2]

lol, calling Infra a liar?
Gigabyte should have gotten it right the first time, simple as that.

They even had an excuse about the board following AMD's specifications "very tightly" yet the ASUS Crosshair V became the reference board for the platform.

[I.nfraR.ed]

I don't see where he calls me a liar. Let's not turn this thread into an argue.

I haven't played with first revision of UD3 but it was actually the best of the line, while UD7 failing hard.
Then they added LLC in the bios in the second revision (1.1?), but it was not working properly. It was finally fixed in the next revision.

[The Stilt]

I can assure you there are no motherboards which would exceed the maximum / minimum limits specified for droop on AMD cpus.

An example:

AMD FX-8350 - 125W TDP (Infra FR, Group A)

TDC = 110A
VID = 1.40000V (for example)

If the AMD specifications are followed:

Maximum allowed voltage at load resulting 110A current draw: 1.317V ((VID + 60mV) - (IDD * 1.3mOhm))
Typical voltage at load resulting 110A current draw: 1.292V ((VID + 35mV) - (IDD * 1.3mOhm))
Minimum allowed voltage at load resulting 110A current draw: 1.267V ((VID + 10mV) - (IDD * 1.3mOhm))

Most of the ODMs allow controlling the LLC separately.
ASUS for example allows to decrease the LLC in 25% steps (0.0013 * 1, 0.0013 * 0.75, 0.0013 * 0.50, 0.0013 * 0.25, 0.0013 * 0.00).

In case the motherboard is throttling under the load, it has nothing to do with the Apm feature.
Apm is part of AGESA so it acts exactly the same on each and every board.
So there are definitely no "bad revision" boards which would throttle because of Apm.

In case the CPU throttles it is because the VRM is insufficient, poorly designed and therefore throttling.

[BeepBeep2]

I don't see where he calls me a liar. Let's not turn this thread into an argue.

I haven't played with first revision of UD3 but it was actually the best of the line, while UD7 failing hard.
Then they added LLC in the bios in the second revision (1.1?), but it was not working properly. It was finally fixed in the next revision.

Sorry, badly worded. I wasn't claiming he called you a liar, but was asking him if he was calling you a liar. Still, I should not have been so hostile.

I can assure you there are no motherboards which would exceed the maximum / minimum limits specified for droop on AMD cpus.

An example:

AMD FX-8350 - 125W TDP (Infra FR, Group A)

TDC = 110A
VID = 1.40000V (for example)

If the AMD specifications are followed:

Maximum allowed voltage at load resulting 110A current draw: 1.317V ((VID + 60mV) - (IDD * 1.3mOhm))
Typical voltage at load resulting 110A current draw: 1.292V ((VID + 35mV) - (IDD * 1.3mOhm))
Minimum allowed voltage at load resulting 110A current draw: 1.267V ((VID + 10mV) - (IDD * 1.3mOhm))

Most of the ODMs allow controlling the LLC separately.
ASUS for example allows to decrease the LLC in 25% steps (0.0013 * 1, 0.0013 * 0.75, 0.0013 * 0.50, 0.0013 * 0.25, 0.0013 * 0.00).

In case the motherboard is throttling under the load, it has nothing to do with the Apm feature.
Apm is part of AGESA so it acts exactly the same on each and every board.
So there are definitely no "bad revision" boards which would throttle because of Apm.

In case the CPU throttles it is because the VRM is insufficient, poorly designed and therefore throttling.

Thanks for the explanation. 1.267v is pretty low for 1.40000V VID though, but interesting that it is still acceptable by specification.

[AlleyViper]

I haven't seen comments on the M5A99FX PRO R2.0, which was sort of a late release, but seems close to the UD3 price. Are those doing well with OC'd FX8Ks under load?