Gigabyte X299 Aorus Gaming 7 Motherboard Review

[Shubham Deshpande]

Gigabyte X299 Aorus Gaming 7 vs MSI x299 Gaming M7 ack
Which 1 is better?

 

[none12345]

"Does anybody know if it is "OK" to put your graphics card in any of the 5 PCIe slots on the motherboard."

As a general rule, the answer is NO. But, the general answer is it depends.

Mainly it depends on how many PCI lanes a cpu supports.

For instance on mainstream intel or amd, thats 16 lanes going to the gpu slots. Often times the primary slot will be the only one with an electrical x16 connection. If it supports sli, the 2nd one is usually x8 electrically. Tho this can vary by motherboard. The difference between x16 and x8 is very small, you probably wouldnt notice it but its there.

If there are 3 x16 slots on those motherboards. The 3rd is almost certainly connected to the chipset, instead of the cpu, and will be a good bit slower then the other 2. It would be fine for compute, but not fine for gaming.

On these x299 boards. They have to support 16 28 and 44 lanes going to the cpu. Because of that, the wiring is going to be a mess. The only slot you can be sure of is slot one likely has a x16 electrical connection. Slot 2 may or may not, but its probably at least x8, after that anything goes.

My gut would guess that a lot of the boards have 2 slots that can be electrically x16(if you have the right cpu installed), with 2 more that are x8 electrically no matter the cpu.

If you want to be sure, download the maunal for a board you are looking at, and see what it says.

 

[Crashman]

"Does anybody know if it is "OK" to put your graphics card in any of the 5 PCIe slots on the motherboard."

As a general rule, the answer is NO. But, the general answer is it depends.

Mainly it depends on how many PCI lanes a cpu supports.

For instance on mainstream intel or amd, thats 16 lanes going to the gpu slots. Often times the primary slot will be the only one with an electrical x16 connection. If it supports sli, the 2nd one is usually x8 electrically. Tho this can vary by motherboard. The difference between x16 and x8 is very small, you probably wouldnt notice it but its there.

If there are 3 x16 slots on those motherboards. The 3rd is almost certainly connected to the chipset, instead of the cpu, and will be a good bit slower then the other 2. It would be fine for compute, but not fine for gaming.

On these x299 boards. They have to support 16 28 and 44 lanes going to the cpu. Because of that, the wiring is going to be a mess. The only slot you can be sure of is slot one likely has a x16 electrical connection. Slot 2 may or may not, but its probably at least x8, after that anything goes.

My gut would guess that a lot of the boards have 2 slots that can be electrically x16(if you have the right cpu installed), with 2 more that are x8 electrically no matter the cpu.

If you want to be sure, download the maunal for a board you are looking at, and see what it says.

Or read the review. The information is all there.

 

[none12345]

Heh i just downloaded the mobo manual for this board.

I was wrong, the first pci sliot is only x16 for the 28 and 44 lane cpus. On the 16 lane kabylake-x its only x8 on the primary slot. Wasnt expecting that. Umm....one more reason why kabylake-x is a utterly stupid choice for x299.

 

[Crashman]

Heh i just downloaded the mobo manual for this board.

I was wrong, the first pci sliot is only x16 for the 28 and 44 lane cpus. On the 16 lane kabylake-x its only x8 on the primary slot. Wasnt expecting that. Umm....one more reason why kabylake-x is a utterly stupid choice for x299.

Like i said, it's in the review. Heck, it's in the table at the beginning of the article:


PCIe x16
(5) 3.0 (@44: x16/x4*/x16/x4/x8) (@28: x16/x4*/x8/x4/x0, x8/x4*/x8/x4/x8) (@16: x8/x4*/x4/x4/x0) (*Shares 2nd M.2)

The reason the table isn't more verbose is that it "needed to be standardized for all motherboards".

 

[Nintendork]

Threadripper made intel's X299 completely irrelevant.

 

[daglesj]

RGB and half the junk on that board has no place on the HEDT platform.

Now where is the version that strips off the junk and sells for $275?

 

[daglesj]

I have a feeling that X299 will be a very shortlived platform.

It just looks a real mess. No one has a clear strategy for it.

 

[Nergo Pthycc]

If only the motherboard manufacturers would design RGB "enhancements" resembling the display at the Bellagio I could be impressed with the value. Or perhaps just build a monitor right onto the motherboard. /s

 

[Crashman]

RGB and half the junk on that board has no place on the HEDT platform.

Now where is the version that strips off the junk and sells for $275?

I don't hit the down arrow often, but when I do it's because someone says "half the junk" without having an apparent clue as to what "junk" they're talking about.

So you'd get rid of RGB. Where's the remaining $100 of the "junk" that's going to magically drop the price of this board to $275?

 

[daglesj]

RGB and half the junk on that board has no place on the HEDT platform.

Now where is the version that strips off the junk and sells for $275?

I don't hit the down arrow often, but when I do it's because someone says "half the junk" without having an apparent clue as to what "junk" they're talking about.

So you'd get rid of RGB. Where's the remaining $100 of the "junk" that's going to magically drop the price of this board to $275?

I don't hit the down arrow often, but when I do it's because someone takes something way too literally.

That kind of thing. It's platform that doesn't need to be tacky.

 

[Crashman]

RGB and half the junk on that board has no place on the HEDT platform.

Now where is the version that strips off the junk and sells for $275?

I don't hit the down arrow often, but when I do it's because someone says "half the junk" without having an apparent clue as to what "junk" they're talking about.

So you'd get rid of RGB. Where's the remaining $100 of the "junk" that's going to magically drop the price of this board to $275?

I don't hit the down arrow often, but when I do it's because someone takes something way too literally.

That kind of thing. It's platform that doesn't need to be tacky.

You'd rather start a downvote war than explain your thoughts on the cost of the board? That's fine. We did review the $300 ASRock X299 Taichi, if you're really interested in something with less lighting and lower cost.

 

[eltouristo]

Re: Asus Prime X299-Deluxe was the only sample to stay within Intel’s 140W guideline. I would really like
some more explanation about all this. It seems to imply the Asus Prime Deluxe is using different settings than the
others. Isn't it obvious we'd all like to know what those differences are? ( if they aren't different, then we have a much much larger issue that we all obviously want to know about). Further, since it seems almost safe to assume it is settings being different, I think we'd like to see what happens when all tested board are set the same way as the ASUS. I don't think this is just me, I am very interested in this exact topic, is vital to me to selecting MB. Thanks!

 

[Crashman]

Re: Asus Prime X299-Deluxe was the only sample to stay within Intel’s 140W guideline. I would really like
some more explanation about all this. It seems to imply the Asus Prime Deluxe is using different settings than the
others. Isn't it obvious we'd all like to know what those differences are? ( if they aren't different, then we have a much much larger issue that we all obviously want to know about). Further, since it seems almost safe to assume it is settings being different, I think we'd like to see what happens when all tested board are set the same way as the ASUS. I don't think this is just me, I am very interested in this exact topic, is vital to me to selecting MB. Thanks!

We tried that. I set the power limit on several boards to 140W and the throttling was choppier, caused greater performance losses, and caused XTU to constantly show the Power Limit Throttling flag.

 

[eltouristo]

hmm. thanks. but I thought you could see most vital parameters, and adjust them manually to match. Excuse my unfamiliarity with certain aspects (not all). A 'power limit' seems like a 'global' 'dumb' sort of macro or something. I mean if all the OC setting are same, boards are supplying same power etc, things should look similar. But I could be confused about review goals here. Maybe I should read it again. With the same voltages, same clocks, behavior and power consumption should be similar. I could be totally misunderstanding, but it sound like some boards are just 'throwing higher settings' at things and so consuming more power. This is all really weird, so it is interesting, lol. I'm used to looking at reviews with marginal difference in power consumption and behavior, I mean, marginal compared to the differences stated here. An extra 100w seem like, well, hardly any sort of comparison, to me it seems things must be set quite a bit different. Differences in efficiencies etc between power related parts on MB could never account for that I think. At first is sounds like an Asus is leading. I'm guess something about firmware or something idk. But obviously other MB makers would be all over this sort of thing so maybe this is 'early adopter' difference and Asus is just 'first out of the gate' with better firmware or something idk. You see what I'm grasping at.. maybe you can help more. I'm still totally confused so I will follow over next week see if I can get insight. thanks

 

[Crashman]

hmm. thanks. but I thought you could see most vital parameters, and adjust them manually to match. Excuse my unfamiliarity with certain aspects (not all). A 'power limit' seems like a 'global' 'dumb' sort of macro or something. I mean if all the OC setting are same, boards are supplying same power etc, things should look similar. But I could be confused about review goals here. Maybe I should read it again. With the same voltages, same clocks, behavior and power consumption should be similar. I could be totally misunderstanding, but it sound like some boards are just 'throwing higher settings' at things and so consuming more power. This is all really weird, so it is interesting, lol. I'm used to looking at reviews with marginal difference in power consumption and behavior, I mean, marginal compared to the differences stated here. An extra 100w seem like, well, hardly any sort of comparison, to me it seems things must be set quite a bit different. Differences in efficiencies etc between power related parts on MB could never account for that I think. At first is sounds like an Asus is leading. I'm guess something about firmware or something idk. But obviously other MB makers would be all over this sort of thing so maybe this is 'early adopter' difference and Asus is just 'first out of the gate' with better firmware or something idk. You see what I'm grasping at.. maybe you can help more. I'm still totally confused so I will follow over next week see if I can get insight. thanks

The other boards are just letting the CPU run nearly unrestricted. You'll generally see around 189W of detected TDP with everything but the stock TDP profile enabled, and 239W with other power-saving features disabled. These are just observations. Manual TDP configuration doesn't appear to get us to exactly the same place as the default feature set.

What we're hearing is that Intel told these board makers they'd basically look the other way if those board makers chose not to follow its TDP parameters.

And if you're wondering what that means, well, Intel's stated TDP isn't realistic for the processor's stated performance level. System makers like Dell are told "you need 140W of continuous cooling power" but if they do that, they end up with the CPU hopelessly throttled at full load. Thermal throttling won't even need to occur because of power throttling, so the 140W TDP becomes redundant to the 140W thermal limit.

In order to get around that problem, manufacturers of retail motherboards are mostly setting up their firmware to disregard the default TDP configuration and run default Intel Turbo Boost multiplier instead. Intel XTU reports this as a variable TDP. And there's your discrepancy between 140W an 189W.

Manufacturers are now adding another feature that they had previously used to boost performance even more: "Enhanced Turbo Ratios", aka "Turbo Ratio Enhancement", "Enhanced Multi-Core Performance", etc. Rather than run the default of "45x for 1 or 2 cores loaded, 41x for 3 or 4 cores loaded, 40x for 5 or more cores loaded", the board sets the processor to "45x, regardless of the number of cores loaded". And then you're looking at even more power and heat.

Remember that this is all from observation: Beyond insider statements that Intel has told motherboard manufacturers that it's OK to ignore default TDP, I have no information about how this firmware is being programmed.

 

[eltouristo]

thanks for great explanation ! I think from my point of view, I would mainly be interested in stable benchmarking with measurements of total system watt usage (using same or similar PS and GPU). I'm used to seeing that sort of thing, those are expressions I can relate to, and it tells you much about performance and heat. So if you have a watt meter and have or can do that, that would great to compare the MBs that way. Idk if what XTU is showing is comparable to that sort of 'actual power' testing. I guess that explains my reaction to 140w vs 239w ! I tend to like Gigabyte simply because I've had luck with reliability. But of course for this build if there's a MB that does the same mild OC with far less watts or even no over-watts, vs the others, I will have to go with that one, since heat/power consumption are already such a concern for a x299 system. Thanks!

 

[Crashman]

thanks for great explanation ! I think from my point of view, I would mainly be interested in stable benchmarking with measurements of total system watt usage (using same or similar PS and GPU). I'm used to seeing that sort of thing, those are expressions I can relate to, and it tells you much about performance and heat. So if you have a watt meter and have or can do that, that would great to compare the MBs that way. Idk if what XTU is showing is comparable to that sort of 'actual power' testing. I guess that explains my reaction to 140w vs 239w ! I tend to like Gigabyte simply because I've had luck with reliability. But of course for this build if there's a MB that does the same mild OC with far less watts or even no over-watts, vs the others, I will have to go with that one, since heat/power consumption are already such a concern for a x299 system. Thanks!

 

[eltouristo]

! [/quotemsg]

[/quotemsg]

this looks like actual system power measurement, which is great.
Would you be able to add some of the basic clocks and voltage info that goes with each other these?
Because it seems they must be set different.
The basic bench score number there would be nice too, though idk if Prime is really a bench, more like a stress test I think.
I thought I saw where they the ones with higher power draw weren't really benching much higher,
not enough to be worth the power. It's seem keeping the power lower should be without much drawbacks really,
but knowing how to be sure you are accomplishing that, or how perf different it might make, seems like really relevant info. Thanks!

 

[Crashman]

this looks like actual system power measurement, which is great.
Would you be able to add some of the basic clocks and voltage info that goes with each other these?
Because it seems they must be set different.
The basic bench score number there would be nice too, though idk if Prime is really a bench, more like a stress test I think.
I thought I saw where they the ones with higher power draw weren't really benching much higher,
not enough to be worth the power. It's seem keeping the power lower should be without much drawbacks really,
but knowing how to be sure you are accomplishing that, or how perf different it might make, seems like really relevant info. Thanks!

Now you're understanding why both the XTU screenshots and the power measurement chart are there no?

As the article states, we enable all of the power saving features before testing performance and power levels, so that performance can be fairly and accurately compared to power levels across various motherboard models. The XTU images show the frequency that results from different manufacturers attempting to bypass Intel's 140W TDP by different levels: Some clocked the CPU down as low as 3.5 GHz. The XTU images show what the motherboard has altered its CPU settings to when fed a full AVX load.

 

[merkela958]

i loved your articles about how GIGABYTE X299 AORUS Gaming 7 , even I have posted one blog on the same: https://techgigi.com/gigabyte-x299-aorus-gaming-7/ . hope you will also like my info too...
you are my inspiration, keep about keep updating your blogs ....

 

[merkela958]

i loved your articles about how GIGABYTE X299 AORUS Gaming 7 , even I have posted one blog on the same: https://techgigi.com/gigabyte-x299-aorus-gaming-7/ . hope you will also like my info too...
you are my inspiration, keep about keep updating your blogs ....