It seems our wait for custom editions of AMD's RX Vega graphics cards is coming to an end. "Better late than never" is what they always say; however, AMD and its AIB partners have to know that this kind of wait can sap customer enthusiasm for a product. It's not enough that customers waited around two years for Vega to come to fruition; we've also had to wait some additional months (not weeks), for an actual custom-design graphics card. Vega's exotic design with HBM2 memory means that these graphics cards' availability would fall prey not only to Vega GPU yields, but also to HBM2 memory availability.

Additionally, Vega has been vulnerable to packaging of HBM2 and the GPU as well, with various factories providing different levels of quality in the finished product. This introduced some unexpected variance in the finished products - making the creation of cooling designs that could cope with all the design discrepancies more difficult.

NVIDIA's decision to prohibit AIB partners from factory overclocking their GTX 1070 Ti graphics is a decision that may not sit well with hardware enthusiasts, but it's one that makes the most sense from a business perspective. As it is, it's expected that overclocked versions of the GTX 1070 Ti graphics cards, which have come to offer a more compelling NVIDIA option against AMD's Vega 56 graphics card, will offer performance that is by all accounts comparable to that offered by its more expensive (and older) GTX 1080 graphics cards. So NVIDIA is just ensuring that its introduction of the GTX 1070 Ti with its $459 MSRP doesn't cannibalize sales of their higher tier graphics card through higher out-of-the-box performance via factory overclocking.

Hot from the semiconductor presses, but even hotter from our very own reviews of Intel's latest-generation Coffee Lake CPUs, Intel's 8700K, unlocked six-core, 12-threaded processor has already been overclocked up to 7.45 GHz under extreme cooling. The feat, achieved by HWBOT user Kovan Yang, currently stands in first place of 8700 K processors, and is in sixth-place in overall CPU frequency ranking - which marks the first time in a while an Intel Core branded processor has achieved such a feat.

The overclock was achieved in a system that was configured with MSI's Z370 Godlike Gaming motherboard, unspecified DDR4 memory, an NVIDIA 8400 GS graphics card. The INtel 8700K was configured with a multiplier of 73x, and the base BUS speed was increased to 101 MHz. Expect more daring overclocks with Intel's latest family of Core processors in the future - der8auer has achieved a speed of 7.3 GHz on the same processor already, but more users and big overclock players are sure to make their overclocking feats known. In our own review, our very own W1zzard found that Intel's latest 8700K was the fastest, more future-proof consumer-level, non-HEDT Intel processor, and that it can be easily overclocked on air to 5 GHz.

Ahead of the 5th October reviews NDA, Lab501 posted their review of the Core i7-8700K six-core processor using samples not provided by Intel, paired with an Aorus Z370 Ultra Gaming motherboard. The tests reveal that the i7-8700K trades blows with the Ryzen 7 1800X in multi-threaded tests, despite two fewer cores, and has a clear leadership in single-threaded tests. It also reveals that the i7-8700K may not be as pricier than the i7-7700K as previously thought. Interestingly, the i7-8700K also spells trouble for "Skylake-X" Core i7 SKUs such as the i7-7800X and i7-7820X, as it offers multi-threaded performance in proximity to them, while being cheaper overall.

The Core i7-8700K is able to sustain its Turbo Boost frequencies of 4.20 GHz better than Intel's other Core X HEDT chips, which translates into higher gaming performance. The tests reveal that today's games still don't need six cores, and on the merit of high sustained clock speeds alone, the i7-8700K is shaping up to be among the fastest processors you can choose for gaming PC builds. Lab501 also got the i7-8700K to overclock to 5.1 GHz with relative ease. The chip runs feisty hot at overclocked speeds, but rewards with HEDT-like performance. Find other interesting findings of Lab501 in the source link below.

While we're still waiting to see AMD's Vega graphics cards undergo a proper custom treatment from the company's AIB partners (MSI included), Micro-Star International has announced at the Tokyo Game Show that they'll be launching yet another version of NVIDIA's GTX 1080 Ti graphics card. This new design revision picks up the triple fan design that has been the staple of MSI's Lightning series of graphics cards, and applies it to the Gaming X brand. The MSI GTX 1080 Ti Gaming X Trio will thus feature 3x of the company's Torx 2.0 fans. The card features a 2.5-slot design that exhausts the hot air to the inside of your case (a minor inconvenience, since MSI's graphics cooling designs are generally considered some of the best out there), and features MSI's Mystic Light RGB tech.

G.SKILL International Enterprise Co., Ltd., the world's leading manufacturer of extreme performance memory and gaming peripherals, is excited to announce a new extreme speed DDR4 memory kits at DDR4-4600MHz CL19! Built with hand-selected, high-quality Samsung B-die IC components, this new addition to the flagship Trident Z series will be available in two color variations: silver aluminum body with white accent bar and black aluminum body with black accent bar.

Redditor "callingthewolf" has posted what is an awe-inspiring result for AMD's Ryzen Threadripper 1950X (that's an interesting username for sure; let's hope that's the only similarity to the boy who cried wolf.) The 16-core, 32-thread processor stands as the likely taker for the HEDT performance crown (at least until Intel's 14-core plus HEDT CPUs make their debut on the X299 platform.) With that many cores, highly thread-aware applications naturally look to see tremendous increases in performance from any frequency increase. In this case, the 1950X's base 3.4 GHz were upped to a whopping 4.0 GHz (@ 1.25 V core) and 4.1 GHz (at 1.4 V core; personally, I'd stick with the 4.0 GHz and call it a day.)

The feat was achieved under a Thermaltake Water 3.0 liquid cooler, on a non-specified ASRock motherboard with all DIMM channels populated with 8 x 8 GB 3066 MHz DIMMs. At 4.0 GHz, the Threadripper 1950X achieves a 3337 points score on Cinebench R15. And at 4.1GHz, the big chip that can (we can't really call it small now can we?) manages to score 58391 points in Geekbench 3. While those scores are certainly impressive, I would just like to point out the fact that this is a 16-core CPU that overclocks as well as (and in some cases, even better than) AMD's 8-core Ryzen 7 CPUs. The frequency potential of this Threadripper part is in the same ballpark of AMD's 8-core dies, which speaks to either an architecture limit or a manufacturing one at around 4 GHz. The Threadripper 1950X is, by all measurements, an impressively "glued together" piece of silicon.

AMD did not just announce retail availability on Ryzen Threadripper today, they also had some on-site and arranged for a fun LN2 overclocking event as part of Capsaicin SIGGRAPH 2017. As always, such events are to give day one estimates on the maximum performance potential of the silicon which in turn guides end users and board partners alike on the worst case scenarios as far as power draw and cooling requirements go.

Monstru from Lab501 was kind enough to share a couple of pictures of the actual event with us while AMD followed up with a Cinebench R15 screenshot as seen below. All 16 cores of the Ryzen Threadripper 1950X were overclocked to 5.2 GHz with a x52 multiplier on a standard 100 MHz bus speed. Core VID from CPU-Z is not trustworthy at these temperatures, so presumably it was more in the range of 1.6 V than 1.16 V. They did have DDR4 RAM in quad channel but at the JEDEC base of 2133 MHz to get as high a CPU frequency without the IMC being a factor. The Cinebench R15 score of 4122 cb is very impressive, given the previous high score for a 16-core CPU was 2867 cb, and it took a 28 core CPU to beat this score before. Sure, the days of high core count overclockable CPUs is only coming now but it goes to show where we were before AMD and Intel both decided to go big this generation.After the break we have some photographs of X399 motherboards from various manufacturers, so be sure to take a look.

Over at Techspot, Steven Walton managed to get a hold of Intel's new six-core, 12-thread Core i7-7800X CPU, and chose to take it for a spin over a levy of gaming benchmarks. The results don't bode particularly well for Intel's new top i7 offering, though: it is soundly beat by its smaller, svelter brother in virtually all gaming tasks.

Out-of-the-box results are somewhat in line with what we would expect: the Core i7-7700K does bring about a base clock increased by 700 MHz compared to the i7-7800X (4.2 GHz vs 3.5 GHz), and has a higher boost clock to boot (4.5 GHz vs 4 GHz.) And as we've seen over and over again, including with Intel rival AMD's Ryzen offerings, frequency usually trumps core count when it comes to performance when applications are exposed more than four cores. And this leads to Walton's results: the Core i7 7700K is still king in pure FPS terms, coming in with a much more attractive proposition than the 7800X in both minimum and maximum FPS, as well as power consumption.

We have some updated information on the X299 Platforms VRM issues from the same overclocker who initially discovered the issue, renowned overclocker der8auer. In an updated YouTube video, der8auer first updated his viewers with new information on his testing techniques, and basically concluded that all issues initially detected (throttling included) are still is an issue even after extensive testing, only in some instances it is difficult to detect not only if you are throttling, but even specifics such as what precisely is throttling. He goes into extensive detail, but a brief summary of the videos main points can be found below for your consumption.

A new DDR4 overclocking record was achieved by Australian overclocker "newlife", breathing new life towards Ryzen's memory frequency support. Don't just count your fireworks right now, though: while impressive the result came with some caveats in the form of the user's Ryzen 5 1400 CPU: it was down-locked to a paltry, performance-murdering 800 MHz.

After shipping with what could be considered by some as broken DDR4 memory support, AMD's Ryzen platform has in the meantime received the proper amount of care such a pivoting product for AMD should. A series of AGESA updates which improved AMD's performance in gaming, as well as DDR4 memory support have been under distribution since the platform's launch, and those updates have surely worked towards achieving this record today. The score was achieved with a single 8 GB G.Skill Trident Z E-die memory kit, which is usually rated for 3600 MHz frequencies (F4-3600C17-4GTZ), using 18-20-20-58-93-1 timings. This module was seated on a GIGABYTE AORUS AX370-Gaming K7 motherboard (F4 BIOS).

At Computex 2017, Galaxy showcased two of its most premium graphics cards based on NVIDIA's GTX 1080 Ti chip. The GTX 1080 Ti OC Lab Edition features a Hall of Fame branded waterblock of appealing design, even if the colored power cables on the left of the unit do break the pleasing color scheme. The card features 3x 8-pin power connectors, for those of you who like to run wild with your overclocks. From the box, this card features a base 1569 MHz clock and a 1683 MHz Boost.

The second graphics card showcased by the company is the GTX 1080 Ti HOF Limited Edition, which makes use of the same 3x 8-pin power connectors, but sheds the waterblock for a triple-fan cooling solution. Strangely, this card features higher out-of-the-box clocks than the waterblock-equipped version, at 1645 MHz base and 1759 MHz boost. There is a LUMIN X branding on the cooler shroud, which indicates the usage of Galaxy's lighting system. It also features an LCD screen for graphics card status information, like operating temperature and current clocks and voltages.

Remember that EVGA card to end all cards? Well, EVGA has now officially announced it. Built in collaboration with overclocker K|ngp|n (Vince Lucido), the GeForce GTX 1080 Ti K|ngp|n Edition Graphics Card sports a cooler that builds upon the iCX technology innovations EVGA has been moving towards. In addition, the company has redoubled efforts towards packing the highest-grade electrical components in this graphics card for extreme overclocking methods, even going so far as to include support (via a connector) to external overclocking modules such as EVBot.

In addition to the X299 Aorus Gaming 9, GIGABYTE showed off its Z270X Aorus Gaming 9, its flagship socket LGA1151 motherboard based on the current Z270 Express chipset. With its nomenclature, GIGABYTE appears to have relegated "Aorus" to a brand-extension, rather than an overarching brand by itself. The Z270X Aorus Gaming 9 is packed to the brim with features. The board draws power from a combination of 24-pin ATX and 8-pin EPS power connectors; conditioning it for the CPU with a massive 24-phase VRM, with a liquid-cooling ready heatsink. This board features a PLX PEX8747 PCI-Express gen 3.0 x48 bridge chip, which takes in the x16 PEG link from the CPU, and puts out four x16 slots (x16/NC/x16/NC or x16/NC/x8/x8 or x8/x8/x8/x8), with support for 3-way and 4-way SLI and CrossFireX.

Storage options are aplenty, with two 32 Gb/s M.2 slots with NVMe RAID and Optane support, two 32 Gb/s U.2 ports, two 16 Gb/s SATA-Express ports, and ten SATA 6 Gb/s ports. This board features the same ultra high-grade AMP-UP onboard audio solution as the X299 Gaming 9, with a 127 dBA SNR CODEC made by ESS, the highest-grade audio capacitors money can buy, and three user-replaceable OPAMPs. The board will feature a plethora of CPU and memory overclocking features, and GIGABYTE's RGB Fusion LED lighting system. It goes on sale some time in June.

Reports around the web (and posts on Intel's forums) speak in hushed, strained and horrified voices at how some users with Intel's Core i7-7700 processors are seeing strangely random temperature spikes on their processors, which prompts their cooling solutions to spin to the rescue. The report only mentions Intel's 7700 (non-K) processor; though it would seem this issue is more prone to happen with the K version of the processor, according to Intel's forums.

Apparently, some users are seeing temperature spikes that reach as high as as high as 90°C (out of a recommended 100ºC.) Some users even go as far as admitting to have replaced Intel's fabled TIM, and running the CPU under a water cooling solution, only to find those temperature spikes still happening - and their cooling solutions rev up in response. "My own chip suffers from it, (without any overclocking) which is quite an annoyance," a user wrote. "This despite a delid modification and a proper water loop, resulting in the fans ramping up and down very frequently, and the temperature appearing to frequently spike near the danger zone." Intel, naturally, deployed a sanitized response, saying that "the reported behavior of the 7th Generation Intel Core i7-7700K Processor, showing momentary temperature changes from the idle temperature, is normal while completing a task (like opening a browser or an application or a program)." Business talk all the way, but to be honest, we don't even know if there is a real problem here, though there are so pretty interesting OCCT graphs being posted on the forum page. What do you say? Any of our users have seen similar issues?

Gigabyte has thrown its hat on the RX 550 line of graphics cards, offering two 2 GB versions of the cards. These sport slight overclocks on their core clock speed, at 1,219MHz (for the OC 2G version) and 1,195MHz (for the D5 2G version.) Like all other RX 550, these carry a 128-bit bus and lack any auxiliary power connectors.

Being entry-level, IGP-substitute cards does not mean AIBs can skimp on cooling - especially not considering these graphics cards now carry more performance (and higher TDP) than some aeons-old enthusiast-level GPUs. As such, these include Gigabyte's Windforce cooler with a patented Blade Fan design and 3D active fan functionality. The company claims an air flow improvement of 23% over traditional fans due to the 3D stripe curve on the fan surface. The semi-passive feature, which is something most AIBs now include in their designs (even if these do somewhat impact the longevity of the fans, due to higher pressure on their mechanisms whenever they start spinning again) allows the fans to remain off at lower temperatures and spin when the GPU is under heavy load. Both cards feature Gigabyte's Ultra Durable construction, which includes solid capacitors and metals chokes. As for software and user control, Gigabyte is bundling the Aorus Graphics Engine software utility with both cards, allowing for one-click overclocking as well as the ability to control clock speeds, voltage, power target, and fan profiles. The Radeon RX 550 D5 2G and the Radeon RX 550 Gaming OC 2G are available now for $80 and $90, respectively.

AIB partner for NVIDIA KFA2 is renowned for the design (often polarizing) of its graphics cards, and the company seems to carry white quite close to its heart (something I don't have a problem with, actually.) Now, the company has seemingly confirmed incoming retail availability of NVIDIA's leaked GT 1030 graphics cards, with an EXOC edition of the card in question.

The EXOC white edition by KFA2 is factory overclocked out of the box, and the packaging confirms its a 2GB GDDR5 model with 64-bit memory bus. According to the source, El Chapuzas Informatico, this card is equipped with a 16nm GP108 GPU with 384 CUDA cores (not the 512 we previously reported.) This makes sense, however, as this means NVIDIA can easily carve a GT 1040 SKU from the supposed 512 CUDA-cores base design of the GP108 chip. The base clock for the KFA2 GT 1030 EXOC is 1252 MHz, with a 1506 MHz boost clock. This card is expected to go on sale for around 80€ ($87 direct conversion, but more likely a $69 price-tag.)

This here is probably the most beautiful AIB partner card I've ever seen (and beauty is in the eye of the beholder, so you are naturally free to disagree with me.) The pristine color scheme and subdued backplate design really stand out for me, as does the white PCB. Nitpicking, I hoped the power cables would not be color coded as they are, but I guess that would invite disaster.

User TonybonJoby in our own forums has successfully flashed his XFX RX 480 graphics card with the BIOS from a Sapphire RX 580 Limited Edition (the one that runs at 1411 MHz Boost clocks, yes.) Having obtained the Sapphire's BIOS right here on TPU, he then flashed it onto his graphics card (which possesses a dual-BIOS setup; this is an important point which you should consider, as it gives you an extra safety net should anything go wrong) through ATIFlash. The newly-christened RX 580 thus smiles for the screenshot, with a stock clock of 1411 MHz, higher than most overclocks possible with the RX 480 cards, probably due to increased voltages on the BIOS level. The user then tested the card on The Witcher 3 and Furmark, with no problems having been reported. Just remember to back-up your BIOS with GPU-Z and make sure to peruse our forums for some details on this flashing process before you get the proverbial grease on your elbows.

Essentially, this may allow you to bypass some artificial overclocking limitation with your graphics card, probably by increased voltages on different power states of the card. You should do this at your own risk, and remember, the only guaranteed way of getting an RX 580 is... you guessed it, buying an RX 580. However, this might also give you an extra performance boost, and free performance is always good, right?

Vendor lists for RX 500 series cards keep popping up, and this time, it's Sapphire and MSI's time. And it would seem that Sapphire has just seen the entirety of its RX 500 series lineup leaked (sans the still absent in battle RX 560.) Apparently, Sapphire will launch a new PULSE line of graphics cards, in addition to its already known NITRO series. This new PULSE line of graphics cards will likely carry previous-gen Polaris 10 chips, judging from the difference in pricing between the top of the line RX 580 PULSE (20G) model and its NITRO (40G) counterpart: a 40€ premium can't really justify a differentiation in overclocking alone. A similar situation is seen in regards to the RX 570 cards, with a NITRO-branded, 8 GB RX 570 (40G) being priced higher than a 4 GB, PULSE-branded RX 570. Looking at the model numbers, it would seem differentiation between the Polaris 10 chips and the Polaris 20 XTX and XL is done by the last characters in the product number, with the "40G" products carrying a hefty premium over the "20G" parts.

If the PULSE series are based on the Polaris 10 chips, and the NITRO are based on the newer, freshly confirmed Polaris 20 XTX, the expected difference in clock speeds (with overclocked variants of the RX 500 reaching 1500 MHz) and the newer, as-of-yet-unconfirmed LPP fabrication process would go a long way towards justifying such a premium. This could speak for an approach on clock-speeds towards differentiating the multiple RX 580 price-points, akin to the 7970's GHz Edition - likely, top-of-the line Polaris 20 XTX and XL chips will board higher-tier graphics cards, marketed at exceedingly high clock-speeds.

G.SKILL International Enterprise Co., Ltd., the world's leading manufacturer of extreme performance memory and gaming peripherals, is thrilled to release new ultra-high capacity Trident Z RGB DDR4 memory kits based on 16 GB modules designed for Intel X99 and Z270 platforms.

Since the widely-acclaimed launch of Trident Z RGB memory series in December 2016, G.SKILL focused heavily on bringing high performance DDR4 16 GB RGB modules to the Intel X99 platform. Now, the Trident Z RGB is available in full sets of 8 modules for a total of 64 GB (8 GB x8) or 128 GB (16 GB x8) capacity, along with lighting control software support for the Intel X99 platform from most major motherboard vendors.

Even though the NDA still isn't up on AMD's second volley of Ryzen-based CPUs, some lucky buyers are already running some of the upcoming Ryzen 5 processors after some sellers jumped the gun. Now, a YouTube video by user "Santiago Santiago." is making the rounds in which he compares gaming performance between the Ryzen 5 1400 (4-core, 8-thread part @ 3.2 GHz base, 3.4 GHz boost), Intel's i5 7400 (4-cores @ 3.0 GHz base, 3.5 GHz boost), and the Pentium G4560, a Kaby Lake dual-core CPU with Hyper Threading @ 3.5 GHz base clocks. The user even snapped a picture proving he has his hands on this chip.

MSI's top of the line take on the GTX 1080 Ti - The GAMING X - has just been detailed by the company, building upon the previously-released teasers. The GAMING X features MSI's two-and-a-half slot TWIN FROZR VI dual-fan cooling solution with Torx 2.0 fans, which have the ability to completely turn off in low-load scenarios through their Zero Frozr feature. The card comes with a custom PCB, equipped with dual 8-pin power connectors and 8+2 phase design.

Overclocking prodigy k|ngp|n has managed what some thought impossible: he pushed a GeForce GTX 1080 Ti over the 3 GHz barrier. And this was done using a modded Founder's Edition card, no less, so chances are some AIB cards will be able to achieve even better overclocks.

The massive overclock on the core was accompanied by a massive memory overclocking as well, which pushed the graphics card to over 500 GB/s of bandwidth. At the time, it isn't known whether the overclock was stable enough for benchmarking - but if it were, this would surely be the fastest consumer-grade GPU on the planet. The overclock was confirmed using TechPowerUp's own GPU-Z.

Power users sometimes really go the extra mile towards achieving the best performance on their hardware. And sometimes, this process includes delidding, as in, removing the processor's Integrated Heatspreader (IHS). This would allow for users to sometimes replace less than perfect TIM (Thermal Interface Material) companies use, achieving lower operating temperatures, and possibly even higher overclocks.

Well, you really shouldn't try to do so with AMD's Ryzen 7. The reason: attempting to delid said processors cost overclocking genius der8auer a grand total of 3 (three!) Ryzen 7 samples before he managed to do it without damaging the processor. This happens because contrary to other CPUs, AMD's Ryzen 7 IHS comes soldered to the chip, which obviously increases difficulty and risk of such a delidding process. Apparently, AMD did a pretty good job with the thermal interfaces of Ryzen 7 anyway - der8auer achieved only a 2ºC decrease in operating temperatures on the delidded Ryzen sample. Long story short: maybe it's not worth it. Especially if your cooling solution of choice isn't able to achieve proper contact with the CPU after the process. You can see a video of the direct cooling test, after the break.