TR Forums

Hey all, I'm looking for an x299 board, and from what I can tell the Asrock seems like the one with the best features/price.

https://www.newegg.com/Product/Product. ... gnorebbr=1

I haven't had great experiences with Asrock products in the past, which gives me pause, I wanted to get your take, a larger sample size if you will.

Thanks in advance!

I'm still rocking my AsRock Z77 Pro4-M and haven't had any issues with it. Sample size = 1.

What's the plan - If you're going for a high-overclock and gaming build, I've been very impressed with the stability of my recent Asrock boards. I know they went through a rough patch several years ago but I've seen nothing recently that is bad, and they do tend to offer the more niche configurations.

If you're building a workstation I'd stay away from the tweaker-heavy gaming boards. You'll likely get a better long-term stability and less BIOS quirks if you stick to a pedestrian model from Gigabyte or Asus. Asus will be way more expensive for no justifiable reason, but their BIOSes are slightly nicer to work in than Gigabyte's. The X299-UD4 looks like a no-frills option where the money is spent on the board components rather than silly plastic bling and coloured lights. If you plan to overclock then the biggest problem with x299 is VRM cooling, by the sounds of it. Get the board with the biggest heatsinks around the socket!

DPete27 wrote:

I'm still rocking my AsRock Z77 Pro4-M and haven't had any issues with it. Sample size = 1.

Sample size = 2 (My Asrock Z77 Pro4-M is also fine).

Actually, we don't tend to buy Asrock at work that often but there are a few weird setups where more traditional vendors don't offer anything (like mATX s2011 boards with 128GB support, or mITX with 8 SATA ports). Sample size ~25 boards and 4 models and we've had no issues with Asrock. In saying that, I think I'd need a sample size of closer to 100 to draw anything meaningful from it.

Here's my main Danger Will Robinson warning about that board: I only see a single 8-pin connector for the CPU power.

Now, that doesn't mean it won't work, it just means that this board is more likely to experience issues taking a large chip like the 7900X or larger variants of Skylake X up to higher overclocked speeds.
Related to that but harder for me to see just based on that listing are things like VRM cooling on the board.

If you just want to drop in a Skylake X (even a 7900X) and run it at stock or with a very mild OC you are fine, but just a caution based on what we've seen around the web.

Yeah, I was wondering what made the OP select that specific board over any other X299 board. Granted these products are a little outside my comfort zone, but a mobo is a mobo. They're priced on features. So what feature does the Fatal1ty Po have that makes it worth the extra.... $130 compared to the the Fatal1ty Gaming per se. These are just questions for the OP to consider. I'd rather not make recommendations based on zero knowledge of the usage needs.

@ Leor - Have you seen this? That's where chuckula's comments on VRMs and # of power pins comes from.

My choice of this board is largely based on this review: http://www.tweaktown.com/reviews/8288/a ... dex11.html

They mention the VRM cooling is very good, and the board is the only one with a N-Base T 10g controller that can also go 2.5 and 5. I hadn't noticed the single 8 pin power connector, that was a good catch. I have an i97900x that's been de-lidded and was hoping to push the chip a bit.

Overall I'm finding it very difficult to find a board that I don't have reservations about.

@DPete27 yes, I'm aware of all of the issues with X299, hence my caution.

leor wrote:

My choice of this board is largely based on this review: http://www.tweaktown.com/reviews/8288/a ... dex11.html

They mention the VRM cooling is very good, and the board is the only one with a N-Base T 10g controller that can also go 2.5 and 5. I hadn't noticed the single 8 pin power connector, that was a good catch. I have an i97900x that's been de-lidded and was hoping to push the chip a bit.

Overall I'm finding it very difficult to find a board that I don't have reservations about.

@DPete27 yes, I'm aware of all of the issues with X299, hence my caution.

From the review the VRM side looks pretty solid and stress-testing a 7900X at 4.6GHz looks decent too. Out of curiosity is this a relatively new board or was it available back in June at launch?

I bought this exact board and a 7900x last week. Just got it up and running over the weekend, replaced an Asrock X99 board which had served me very well (read: no issues) during its lifespan. I've also used several Asrock mainstream boards (Z77-Z97) with no issues and good overclocking potential.

I looked at all the X299 boards out there and said the featureset of this one is just the best (10G NIC + 10 SATA ports and decent VRM temps on a BEEFY 13 phase VRM). In my system the VRM heatsink is directly underneath a radiator fan so I'm not too worried about temps, and since this is a 24x7 media server I'm not going to push OCs to the limits (shooting for 4.5GHz or so under water).

techguy wrote:

I bought this exact board and a 7900x last week. Just got it up and running over the weekend, replaced an Asrock X99 board which had served me very well (read: no issues) during its lifespan. I've also used several Asrock mainstream boards (Z77-Z97) with no issues and good overclocking potential.

I looked at all the X299 boards out there and said the featureset of this one is just the best (10G NIC + 10 SATA ports and decent VRM temps on a BEEFY 13 phase VRM). In my system the VRM heatsink is directly underneath a radiator fan so I'm not too worried about temps, and since this is a 24x7 media server I'm not going to push OCs to the limits (shooting for 4.5GHz or so under water).

thanks, this is helpful, how long have you been running the system? I've had a few Asrock boards be a bit wonky in terms of quality of bios and driver updates, what are your impressions?

@everyone Should I be concerned about the single 8 pin connector for power on the board?

leor wrote:

@everyone Should I be concerned about the single 8 pin connector for power on the board?

I'll give the best answer by equivocating. If your goal is to put in a 7900X or lower and keep the overclocking in the "normal" range like 4.5 to 4.6 Ghz on a 7900x (bump that up a notch for lower core counts) then sure.

If you want to shoot for 4.8/4.9 on a 7900x, I'd maybe look for a board with two CPU power connectors.

If this is potentially taking one of the big Skylake X parts that are coming out later this year (even as a future upgrade)... I'd be a little more cautious although I suspect you'd be able to run them at stock speeds without an OC.

Considering it's a 7900X, 8-pin CPU power should be fine as long as you've got a good power supply with a well-built connector and a bit of airflow over the area. That said, I would check by hand that it doesn't get ridiculously toasty under heavy load.

I've never paid or specced more than $150 for a mobo and don't feel qualified to comment further on this particular board.

==== ==== ==== ====

ASRock is often my top pick, and I've built probably a half-dozen systems on their boards now. One was Sandy-era, the rest are Z97 or later.

* Reliability is tough to say without bigger sample sizes. Judging by Amazon/Newegg reviews, all brands suck at that and ASRock seems to be about in line with the rest. As far as anecdata, only one of these ASRock builds I've done has run into trouble, but that one that did was the board eventually killing whatever CPU was stuck in it (Z97E-ITX/ac).

* Their firmware interfaces do have occasional glitchiness, but again IME not worse than the others. The interfaces and featuresets themselves are excellent these days. The glitches I've run into are that the mouse would sometimes be unusable in UEFI (seems to be fixed with B350-gen firmware), modifying fan curves occasionally locks the fan controller until the next reboot (only an issue as you're dialling in curves), modifying other settings sometimes re-enables SMT on my B350 board when it's supposed to be disabled, and a bunch of the usual Ryzen memory controller wonkiness.

* Firmware updates have been easy and flawless, and I haven't yet run into a situation where one was really necessary (though I probably will soon for the big Ryzen bug).

* I never use in-OS tools provided by mobo vendors and have no comment on them.

* ASRock did disappoint me on this B350 board (AB350 Gaming-ITX/ac) by using a thermistor under the socket for CPU temp sensing instead of getting it directly from the CPU. It may be a semi-typical AMD thing, but it's still very annoying.

* I buy lots of ASRocks because they tend to be a few bucks cheaper than the competition for the same featureset and don't seem to be notably worse at anything. Given price and feature parity, I'll generally choose the ASRock just because I'm familiar with their firmware interface.

leor wrote:

techguy wrote:

I bought this exact board and a 7900x last week. Just got it up and running over the weekend, replaced an Asrock X99 board which had served me very well (read: no issues) during its lifespan. I've also used several Asrock mainstream boards (Z77-Z97) with no issues and good overclocking potential.

I looked at all the X299 boards out there and said the featureset of this one is just the best (10G NIC + 10 SATA ports and decent VRM temps on a BEEFY 13 phase VRM). In my system the VRM heatsink is directly underneath a radiator fan so I'm not too worried about temps, and since this is a 24x7 media server I'm not going to push OCs to the limits (shooting for 4.5GHz or so under water).

thanks, this is helpful, how long have you been running the system? I've had a few Asrock boards be a bit wonky in terms of quality of bios and driver updates, what are your impressions?

@everyone Should I be concerned about the single 8 pin connector for power on the board?

I've only had it up and running since the weekend, but no issues or concerns thus far. I hesitated on choosing this board or its cousin the Taichi (same board minus 10G NIC) due to the inclusion of only a single 8 pin CPU power connector and the best information I found indicated that none of the other boards have VRMs that are spec'd to provide significantly greater power than this board can without overheating their VRMs. As chuckula and synthtel indicated, maximum CPU power delivery isn't really an issue unless you're OC'ing the snot out of one of these chips. You won't get to that point without sub-ambient cooling or at the very least de-lidding the chip.

Edit: some of the info I used to help make my decision was found at ocn: http://www.overclock.net/t/1632665/inte ... vrm-thread

edit 2: found the actual table of power delivery capabilities - the *only* board with better real world capability is the Rampage VI Apex, a board specifically made for OCing which doesn't appear to be for sale anywhere at this point in time and whose price is unknown, but almost certainly higher than that of the Asrock board in question.

I run ASRock almost exclusively, though I do intermix a few ASUS and Gigabyte boards for diversity.

I have 7 ASRock boards running at home now in my rack that run every generation of the i7's from Gen2 upward.

The systems run 24/7 and the rack temp stays below 97F (typically 95F). I try to keep the systems quiet so I use a max of 3 120mm fans, one of which is on a tower heatsink (usually CM 212's), so the systems don't run the coolest, nor do the VRMs. When they go down for maintenance, power is drained for about a day while I find time finish the upgrades.

Zero problems, never had to RMA a board. I was a bit worried though with a z97 board I had running for about a year. Lots of comments on the Newegg reviews saying that the board works fine until you drain the power, then it never powers on again. Apparently it was a known issue and there was lots of noise about it. I would periodically power that system off and drain the power before the warranty ran out just to make sure. That board eventually wound up in a test bench PC and has been ticking away ever since. The fact it was seemingly such a widespread issue however made me a little worried about the QA and it was the first time I saw such widespread complaints about an ASRock board.

I've yet to upgrade to the x299 but I expect to have a system running in the next few months.

techguy wrote:

I've only had it up and running since the weekend, but no issues or concerns thus far. I hesitated on choosing this board or its cousin the Taichi (same board minus 10G NIC) due to the inclusion of only a single 8 pin CPU power connector and the best information I found indicated that none of the other boards have VRMs that are spec'd to provide significantly greater power than this board can without overheating their VRMs. As chuckula and synthtel indicated, maximum CPU power delivery isn't really an issue unless you're OC'ing the snot out of one of these chips. You won't get to that point without sub-ambient cooling or at the very least de-lidding the chip.

Edit: some of the info I used to help make my decision was found at ocn: http://www.overclock.net/t/1632665/inte ... vrm-thread

edit 2: found the actual table of power delivery capabilities - the *only* board with better real world capability is the Rampage VI Apex, a board specifically made for OCing which doesn't appear to be for sale anywhere at this point in time and whose price is unknown, but almost certainly higher than that of the Asrock board in question.

I had my chip delidded at silicon lottery and they tested it to run stably @ 4.6, 1.20v

I had the thought to bump the voltage a bit and see if it would do 4.8, since I went through the bother of delidding.

leor wrote:

techguy wrote:

I've only had it up and running since the weekend, but no issues or concerns thus far. I hesitated on choosing this board or its cousin the Taichi (same board minus 10G NIC) due to the inclusion of only a single 8 pin CPU power connector and the best information I found indicated that none of the other boards have VRMs that are spec'd to provide significantly greater power than this board can without overheating their VRMs. As chuckula and synthtel indicated, maximum CPU power delivery isn't really an issue unless you're OC'ing the snot out of one of these chips. You won't get to that point without sub-ambient cooling or at the very least de-lidding the chip.

Edit: some of the info I used to help make my decision was found at ocn: http://www.overclock.net/t/1632665/inte ... vrm-thread

edit 2: found the actual table of power delivery capabilities - the *only* board with better real world capability is the Rampage VI Apex, a board specifically made for OCing which doesn't appear to be for sale anywhere at this point in time and whose price is unknown, but almost certainly higher than that of the Asrock board in question.

I had my chip delidded at silicon lottery and they tested it to run stably @ 4.6, 1.20v

I had the thought to bump the voltage a bit and see if it would do 4.8, since I went through the bother of delidding.

4.6 @ 1.2 definitely won't put you over the limits of the 8-pin power connector. 4.8 @ ~ 1.3 will make the connector hot for sure, but provided you have adequate CPU cooling as well as decent airflow in the socket area that should be doable as well. The effect of inadequate power delivery under extreme load exceeding the limit will be one of two things:
1) failure to sustain boost clocks (provided you use turbo boost rather than a fixed clock setting to OC)
2) BSOD/shutdown (if you used fixed clocks/voltage)

That's what happens if the VRMs can't handle it, but any halfway decent VRM design should be able to handle 11.5V input without a care (unless it shuts down on the theory the connector or some such is in danger). 0.5V drop at 360W / 30A total draw would be 15W dissipation in the cable and connector, though. If there's an issue, it isn't with the VRMs.

The failure mode is the connector melting, but for a good connector with a touch of ambient airflow, it just isn't going to happen unless you're pushing truly ludicrous amounts of power through it.