-250 degrees Celsius? Impressive, I do wonder what they had to do with the processor and motherboard to get 7.5 Ghz stable enough for a valid benchmark run too. Will be interesting to see what the normal overclocking capabilities are of the x299, guessing reviews are upcoming of both that and the i9 processors once you can procure some of them what with Intel and sending review samples.

Someone pls explain why they benchmark with old software like 3DMark03, 3DMark06, and Aquamark? I understand they want to compare it with previous results, but isn't it about time to upgrade already? You know, 11 years overdue?

sorry, but anything beyond air/water cooling is irrelevant for 99.9% of all customers (and "real life" use). no mentioning on how many cores either. so far, asus/asrock/msi offer more/better components in the same price bracket, with asrock having one of the best power/vrm quality, and gigabyte willing to cut corners on components after the first batch of reviews are out (at least for the gpus), some LN2 (or similar cooling) is not going to change my mind... show me records when running water/air and its a different story.

If amd or intel could develop low heat cpu's then you could put a freezer unit on and get some good results, Using liquid nitrogen must be fun but dangerous too, These tests show how much more power cpu's have if only they could be cooler.

I do find these record breaking speeds to be of interest, but as fry178 mentioned, 99,9% of us will never come anywhere close to those kind of data crunching numbers with a full core count on regular cooling.

Someone pls explain why they benchmark with old software like 3DMark03, 3DMark06, and Aquamark? I understand they want to compare it with previous results, but isn't it about time to upgrade already? You know, 11 years overdue?

I suspect it was only on one core, the overclock. All of those are single threaded, and very CPU bound nowadays.

gigabyte willing to cut corners on components after the first batch of reviews are out (at least for the gpus).

got a source?

got a source?

Happens occasionally, GPU's and motherboards get a different revision with some tweaks which can either help or worsen the hardware. Article I found is a few years old now but. https://uk.hardware.info/reviews/5835/spot-the-differences-gigabyte-motherboard-revisions-present-markedly-different-test-results Pretty sure it's not just Gigabyte.

Depending on what CPU they used, I find this very believable, and it wouldn't surprise me if some users could get close to 6GHz on consumer-level liquid cooling. Think about it: this motherboard was designed to handle 18 overclockable cores (well... actually 16 cores... but that's still a lot). That is no ordinary task, and certainly not one that Gigabyte can go cheap on. I imagine the VRM phasing is insanely high and the socket is designed with great durability. If they tested with something like a "measly" 8-core i9, reaching an overclock this high is no surprise.

sorry, but anything beyond air/water cooling is irrelevant for 99.9% of all customers (and "real life" use). no mentioning on how many cores either. so far, asus/asrock/msi offer more/better components in the same price bracket, with asrock having one of the best power/vrm quality, and gigabyte willing to cut corners on components after the first batch of reviews are out (at least for the gpus), some LN2 (or similar cooling) is not going to change my mind... show me records when running water/air and its a different story.

mind to show some fact that asus/asrock/msi using better components for their mobos compared to gigabyte? imo mobo nowdays using same parts, solid chockes/caps i never use asrock, but one thing for sure more VRM (ie 12phase) is not always mean better than lower vrm (ie. 8phase), and i dont think what asrock use in their board better than what asus used what important is in design layout and also electrical design that can improve overall performance regarding the OC record, while almost all people wont use LN2 for their own system but getting high record and stable during the test is sort to prove their mobo performance which also can mean "quality" in their board so on air/water system the result should be also good if the cooling work great, as the limitation is not on the mobo but rather on the cooling system, i believe that what gigabyte intend to with their OC promotion

But who cares? Until there's a closed loop liquid helium cooler I can put in my pc and run 24/7, this is completely useless. We've already seen it with the ryzen chips that it means nothing for normal overclocking.

Did they not mention what CPU they used at all? Isn't that relevant?

It was on a 7740K with Liquid Helium on 4c4t cores enabled I believe

Did they not mention what CPU they used at all? Isn't that relevant?

I heard it was a 7740x and all 4 cores were active. Not sure if HT was active though. Edit: damnit Matt beat me.

If Gigabyte X299 can handle 7.5ghz and -250 then i'm sold on the products. 5ghz should be absolutely fine for 24/7 usage. I only use Gigabyte mobos anyway but i'm glad to see the company do well since they gave me a 980G1 for free.

well you do not want ln2 or dry ice , etc etc any cooling solution that goes bellow the room temperature will form moisture on it all over it and droplets so if you put sub room temperature cooling solution on your pc then you have to isolate it really well ! also i think that this kind of extreme conditions those components will not last for long after that no ?((someone that does extreme overclock give me your lights on that one :P))

well you do not want ln2 or dry ice , etc etc any cooling solution that goes bellow the room temperature will form moisture on it all over it and droplets so if you put sub room temperature cooling solution on your pc then you have to isolate it really well ! also i think that this kind of extreme conditions those components will not last for long after that no ?((someone that does extreme overclock give me your lights on that one :P))

I used a Z77X UD3H multiple times under single stage, dice and ln2 and it still works just fine today 😉 And yes it's generally a good idea to insulate the board, I've used eraser, LET and plasti dip to do so, then lay some blue shop towel down I quite enjoy it, it's like drag racing, never understand the half the comments about it on social media etc surely it's extremely obvious it's not for 24/7 lol and is a hobby and source of enjoyment for some 🙂

@buhehe just go to their website and look how many bios updates were released for different vram than the v1.0 and it was always for cheaper chips (samsung to hynix) and not one card i tried was able to run vram as fast/faster than with 1.0. they might not do it anymore, but after experiencing this with Nv 700/900/1000 series, i'm sticking with other brands, as msi also got the better board/cooler design anyway. @slyphnier maybe not on components that much anymore, but definitely on the what the board offers. i havent build a new rig for a few years, but until my recent intel in 2014, i was always able to get the top amd board from asus/asrock, when gigabyte/msi would only offer the next one down for similar price (e.g. asus 990FX-TUF vs UD5, and not the UD7). and i wasnt shopping at bestbuy.. just read the specs for similar priced asus/msi/gigabyte, and compare to asrock. same for online reviews. they all state that the asrocks have some of the best components when it comes to power design/components. at least the ones that took the board "apart" and/or looked at vrm/phases/power delivery under load. gigabyte UD7 doesnt mention any detailed info on parts used for power, besides 6+4 phase and use of digi-design with 10K caps, but tons of info on effin rgb crap or the fan headers. and it doesnt even have 2 nvme slots, uses killer-chip (remember issues either with chips/drivers in the past), and still costs more than the asrock. the asus HERO offers a LN2 button/mode but runs only 10K caps (laughs), asus lists a bit more detail info on parts for power, but still costs 50-100$ more.. the asrock taichi offers 2 nmve slots (vs UD7), has 12k caps, offers double stacked mosfets, lists power chokes as 60A, has 2oz copper, which the others dont offer, and while cheaper than the other 2. and that isnt even the top board for them... dont like the msi looks right now, so i didnt even check those. im not saying the others are bad, but i get "more" for less with asrock, and on their gaming lineup the even offer "gaming" usb-port with 1000Hz pulling.

@slyphnier maybe not on components that much anymore, but definitely on the what the board offers. i havent build a new rig for a few years, but until my recent intel in 2014, i was always able to get the top amd board from asus/asrock, when gigabyte/msi would only offer the next one down for similar price (e.g. asus 990FX-TUF vs UD5, and not the UD7). and i wasnt shopping at bestbuy.. just read the specs for similar priced asus/msi/gigabyte, and compare to asrock. same for online reviews. they all state that the asrocks have some of the best components when it comes to power design/components. at least the ones that took the board "apart" and/or looked at vrm/phases/power delivery under load. gigabyte UD7 doesnt mention any detailed info on parts used for power, besides 6+4 phase and use of digi-design with 10K caps, but tons of info on effin rgb crap or the fan headers. and it doesnt even have 2 nvme slots, uses killer-chip (remember issues either with chips/drivers in the past), and still costs more than the asrock. the asus HERO offers a LN2 button/mode but runs only 10K caps (laughs), asus lists a bit more detail info on parts for power, but still costs 50-100$ more.. the asrock taichi offers 2 nmve slots (vs UD7), has 12k caps, offers double stacked mosfets, lists power chokes as 60A, has 2oz copper, which the others dont offer, and while cheaper than the other 2. and that isnt even the top board for them... dont like the msi looks right now, so i didnt even check those. im not saying the others are bad, but i get "more" for less with asrock, and on their gaming lineup the even offer "gaming" usb-port with 1000Hz pulling.

based from the review i read, like on here guru3d, or anandtech never mention much every single component used on mobo because all maker nowdays using same type, japan caps 10~15k hours rated whose review that reviewed that detail down to the caps ? for asrocks boards based what i heard and read (review) overall they are good reliability-wise asus,asrock&gigabyte are top tiers for mobo-maker ... so picking one of those isnt wrong not for MSI though... not sure if they improve their quality nowdays? but except i get great deal(price), i wont buy their mobo, i got bad luck with their gpus (2 dead already) why u laugh at 10k caps ? u think all boards should use 15k hours caps? between 10k and 12k caps isnt much difference durability wise lets do simple math: if a board using 12k caps but run hot 85C then it last 5.5years... while other board using 10k caps but run cool ~75C then it can last 9years try look those server board, and check they using better caps than consumer mobo... supermicro still using 8k~10k caps from what i know regarding the price, i think it is all about distribution in your country/area for my area - asia(japan) gigabyte pricing most of time lower than asus or asrock, asus usually the highest/premium priced and asrock board availability is kinda limited compared to asus/gigabyte in the end yea, when buying board (and rest of pc parts), its all about price/value and warranty/rma support ...its not about brand/maker

I used a Z77X UD3H multiple times under single stage, dice and ln2 and it still works just fine today 😉 And yes it's generally a good idea to insulate the board, I've used eraser, LET and plasti dip to do so, then lay some blue shop towel down I quite enjoy it, it's like drag racing, never understand the half the comments about it on social media etc surely it's extremely obvious it's not for 24/7 lol and is a hobby and source of enjoyment for some 🙂

well when i told you do not want this kind of cooling solution i meant for 24/7 :P and yes i completely get that it is a sport kind of envying you hehe ... about the life time i had the idea that very often after few hours or benching the components will be worn out say before you force the poor cpu to 7 ghz on -80 -- Celsius to run and validate benches it might be able to do say 4.5ghz on air with 1.25 vcore will need 1.4 vcore after the inhumane torture:P or will not be able too at all , although i have no idea why this was stuck in my head ! The only time i was getting hyped about going completely bananas on hardware was an old q6600 from a friend that he did not wanted so i was planning to get my hands on it and keep pushing it till something breaks ...but at the end he gifted the cpu and mb to someone that really wanted it for his main pc 😥 😥 😥