Click on the competition images to go straight to the competition page, or click here for a more detailed overview at HWBOT.
World Tour 2017 and HWBOT X
Coming soon ...>
Road to Pro 2017
Starts Feb 1, 2018
|GPUPI - 1B||Titan V||k|ngp|n||2sec 559ms|
|Catzilla - 720p||Titan V||k|ngp|n||86421 marks|
|GPUPI - 1B||Titan V||k|ngp|n||1sec 560ms|
|XTU||Core i7 4720HQ||a.a.z||861 marks|
|3DMark - Fire Strike||GeForce GTX 1080 Ti||DStealth||28343 marks|
|3DMark - Fire Strike||GeForce GTX 1080 Ti||yerrihakim||27350 marks|
|XTU||Core i7 8700K||riccardoPolaris||2767 marks|
|3DMark06||Radeon HD 6870||shar00750||37972 marks|
|XTU||Core i7 8700K||dorkyshrew||2753 marks|
|3DMark Vantage - Performance||Radeon HD 6870||shar00750||27862 marks|
Click on the competition images to go straight to the competition page, or click here for a more detailed overview at HWBOT.
Coming soon ...>
Starts Feb 1, 2018
Today we find the GPU Flashback Archive delving into the not so distant past to focus on the NVIDIA 900 series of graphics cards, the first to use NVIDIA’s new Maxwell architecture which had already seen the light day in mobile GPU solutions, an indication of the direction that the company were taking at the time. Let’s take a look at the cards that were launched as part of the 900 Series, the improvements and changes that Maxwell brought and some of the more memorable scores that have been posted on HWBOT.
The first question one may well have regarding the NVIDIA 900 series is simple - what happened to the 800 series? To answer the question fully, you must first look at the direction that NVIDIA was moving at the time. A movement to expand its product offerings in order to compete in the quickly expanding mobile SoC market. The suddenly ubiquity of Android-based smartphones around the globe was fuelled in part by the development of mobile SoCs from Qualcomm, Samsung, Mediatek, Marvell, Allwinner and others. The traditional feature phone was quickly being replaced by smartphones that now required improved multi-core CPU performance, HD display support and, importantly from NVIDIA’s perspective, decent enough graphics processing to actually play 3D games. Intel and NVIDIA were two companies with plenty of R&D and marketing budget who sought to enter a new market to help bolster revenues during an inevitable slow down of desktop PC sales, a traditional cash cow for both.
The GPU Flashback Archive series continues today with a recap of the NVIDIA GeForce 700 series, a series refresh which heralds part two of the Kepler family of GPUs. We can also remember it as a time when NVIDIA launched their first ever GTX Titan card and with it, a new pricing and retail strategy for truly high-end graphics card products. Let’s take a look at the new Kepler architecture GPUs, the cards that were popular with HWBOT members and some of the more memorable scores that have been posted since launch.
The 2011-2013 period of history saw NVIDIA implement a more regular cadence to their high-end product launches and refreshes. One that saw the company launch a new GPU architecture every two years, with new product lines arriving each year. This means deriving two product lines per architecture with an improved version offered the second time out. This is what we saw with Fermi, an architecture whose potential was full realized at the second attempt. With the GeForce 700 series, which arrived proper in May 2013 with the arrival of both the GeForce GTX 780 and GTX 770, we have something different. The new cards arrived using a much bigger version of the Kepler architecture compared to what we saw on the NVIDIA 600 series.
The GPU Flashback Archive arrives today at the NVIDIA 600 series that debuted in Spring of 2012. The new range of cards showcased a new graphics architecture design and the beginning of what we might describe as the Kepler era. Let’s take a peek at the changes that the new design heralded as well as a close up view of on the GeForce GTX 680 card, the most popular 6-series card with HWBOT members historically speaking. Before we look at some notable scores that were made with the GeForce 680, let’s first kick off with an overview of what innovations arrived with the new Kepler architecture.
If we cast our minds back to 2012 we can recall a era when NVIDIA and AMD were virtually neck and neck, with successive graphic card launches from each company swinging the performance crown from side to side. The arrival of Kepler in many ways represents the beginning of the end of the competitive duopoly that is clearly absent today. Kepler helped NVIDIA push ahead of AMD in terms of graphics processor design, creating a performance lead which AMD still finds insurmountable, despite the arrival of their latest Vega-based cards. Let’s take a look at Kepler in a little detail.
This week the GPU Flashback Archive sets its sights on the GeForce 500 series from NVIDIA. Arriving in late 2010, the 500 Series was the second round of graphics cards based on the Fermi architecture which had limped over the line in the previous generation, ostensibly due to fabrication and yield issues. The new flagship GTX 580 arrived with a more polished take on the Fermi design that help NVIDIA combat the threat from AMD and their popular Radeon 5000 and 6000 series cards. As ever, let’s take a look at the new GPU, the new flagship card and a few of the outstanding scores that have been submitted to HWBOT.
To say that the NVIDIA 400 series graphics cards launch was less than smooth, would be a total understatement. The GF100 Fermi architecture GPU in fact arrived six months late with a significant number of cores hacked off. Blame was laid at the door of fabricators TSMC and a 40nm manufacturing process that clearly hadn’t been optimally adapted for NVIDIA’s Fermi, a monster chip boasting 3 billion transistors and a 529mm² die. While cards such as the GTX 480 had actually done well to make NVIDIA competitive in performance terms, the GTX 580 and its GF110 GPU was rather quickly shoved out the door just eight months later as a revised and improved version of the original.
This week in our GPU Flashback Archive series we cast our minds back to a very popular and well loved graphics card series, the GeForce 400 series. NVIDIA launched the GeForce 400 series in March 2010 armed with a new Fermi architecture that it hoped would help it compete with the successful AMD Radeon 5000 series. Let’s look at the new features that Fermi offered, the cards that were popular and the scores that were submitted to HWBOT in this era.
Compared to previous product launches from NVIDIA, the GeForce 400 series launch did not go as smoothly as hoped. September 2009 saw AMD come out with their Radeon 5000 series which made a solid case against NVIDIA 200 series offerings. It would be January before NVIDIA really started wooing tech media with tales of its forthcoming Fermi architecture lineup. It would be March 2010 before tech media actually got their hands on the new cards and several weeks after that before enthusiasts would be able to actually buy one. This was not the typical NVIDIA launch. Reasons for the delay certainly seemed to lie with issues with actual fabrication at TSMC who were not providing the yields expected on their new 40nm process. This was a problem that particularly hurt NVIDIA due to the fact that the new Fermi GPU, the GF100, was actually very large. When the GeForce 400 series finally arrived in the form of the GeForce GTX 480 and GTX 470, by most calculations they were six months late.
It’s fair enough to say that not too many of us will ever actually own an NVIDIA Titan V graphics card, simply because $3,000 USD is a wee bit beyond most people’s budgets. However, that doesn’t mean that we don’t enjoy content produced around the latest and greatest video card ever produced – the voyeur in us is content to watch from a distance. Which is why you may well enjoy this video from Gamers Nexus which gives us a very detailed look at the components used to create the Titan V VRM design, plus a guide on how to perform a Shunt Mod.
The video is hosted by the one and only Buildzoid from Actually Hardcore Overclocking who takes us by the hand on a journey through the entire PCB, scribbling notes to explain that each and every component actually does. He starts with the vCore VRM of course, which in this case is a little odd in that it spans two sides of the GPU. The design does however make sure that the VRM components are as close to the GPU as possible as it means the current has a shorter distance to travel which reduces voltage dropping and improves transient response. The High Bandwidth Memory (HBM) gets its own two phase VRM which just a bit further back.
The VRM itself is described by Buildzoid as an ‘absolute monstrosity’. In fact it is a 16-phase design which uses a voltage controller from Monolithic Power Systems (MPS) that supports 8-phases. The card uses no doublers which it’s basically a massive 8-phase with double the amount of very high-end powerstages and inductors. What you get in the end is a ridiculous amount of potential power delivery.
After the full analysis of the PCB, Buildzoid then moves on to offer a guide regarding Shunt Modding, an easy mod which should in theory give you the chance to pump more power through the GPU. Catch the video from Buildzoid here on the Gamers Nexus YouTube channel.
Steve Burke and Gamers Nexus just went and splurged $3,000 USD on the latest and greatest graphics card from NVIDIA, the Volta-based NVIDIA Titan V. But hey, being a true overclocker and PC enthusiast, Steve just could not resist the temptation to do a tear-down video, removing the entire cooling apparatus to show us what NVIDIA had done at the PCB level. Cheers Steve.
So what is the new Titan card all about? Well, the Titan V of course uses ‘Big Volta’, namely the GV100 graphics processor which you can also find powering the latest Tesla V100 HPC card. The GV100 is a multi-chip module that features a GPU die and three HBM2 memory stacks on the same package. It has 12GB of HBM2 memory that is linked via a 3,072-bit wide memory bus. The GPU has a base clock of 1200MHz boosting by default to a frequency of 1455MHz. Default HBM2 memory is configured at a clock of 850 MHz.
Steve removes the cooler with ease, noting how the card uses the same cooling design that we’ve seen on previous Titan and recent Founder Edition cards. The GV100 GPU is well served in terms of power, being teamed up with a 16-phase VRM. Steve also notes how there is a massive amount of thermal paste used, perhaps not so surprising seeing as the GV100 is a supercomputer chip designed for machine learning and Ai applications and as such has a 815 mm² die, which is pretty much as big as you will ever encounter. Steve goes on to clean up the GPU where you can see the HBM parts of chip package.
You can find the tear-down video from Gamers Nexus here on their YouTube channel.
Today the second ever season of the Cheapaz Chips contest series gets under way over on OC-ESPORTS. As you may recall the idea of the contest is to encourage overclockers to put all of their ingenuity and passion into overclocking entry-level hardware that we wouldn’t cry over too much if it got bricked. Cheapaz Chips Season 2 will run through what remains of 2017, ending on January 15th. It features three stages suited to benching NVIDIA GT 1030 graphics cards, the subject matter of season 2. As an added incentive, GALAX are also offering a GALAX GTX 1080 Ti HoF OC Lab Edition card for the eventual contest winner.
Cheapaz Chips Season 2: December 15th - January 31st 2018 - The Cheapaz Chips Season 2 contest on OC-ESPORTS features a very simple format based around Pascal-architecture NVIDIA GT 1030 cards. The Cheapaz Chips contest has in fact been designed to give overclockers a wonderful excuse to get down to some serious card modding. Graphics card modding can be a daunting task for anyone attempting to do so for the first time however. It takes a steady and assured hand that knows how to solder with unerring accuracy, as well as detailed knowledge of exactly how the card and power delivery design works.
1st Place Wins a GALAX GTX 1080 HoF OC Lab Edition - The fantastic thing about Season 2 of the Cheapaz Chips series is that GALAX are generously offering a GTX 1080 Ti HoF OC Lab Edition card for the winner. The highest scoring overclocker wins a GALAX GTX 1080Ti HOF OC Lab Edition graphics card. In case of a tie, the deciding benchmark is Superposition 1080P Xtreme, then GPUPI 1B, then 3DMark03. Each submission must include a picture of your graphics card in action. The graphics card does not come with international warranty.
Swedish OC legend Rauf has recently been having a very good time benching a pair of NVIDIA GeForce GTX 1080 Ti cards. In fact using GALAX HoF Lab Edition models under LN2, he managed to take down the World Record in the Catzilla 720p benchmark, plus three other Global First Place dual-GPU scores in3DMark Time Spy, 3DMark Fire Strike and Fire Strike Extreme. Let’s take a quick look at those scores and check out his rig.
In the Catzilla 720p benchmark we find a new World record score of 114,843 marks. The score was made with both GPUs pushed at subzero temperatures to 2,300MHz (+55.41%) on the GPU and 1,566MHz (+13.81%) on the graphics memory. The rig was based around a Intel Core i9 7980XE 'Skylake-X' processor clocked at a very tasty 5,500MHz (+111.54%) with a GALAX HoF DDR4 kit configured at 2,000MHz (12-11-11-28). The rig also featured a ASUS ROG Rampage VI Apex motherboard and an 8 Pack Edition Super Flower 2,000W power supply.
The above rig was also used to set a new Global First Place score in 3DMark Time Spy, with the dual-GPU record now standing at 24,389 marks. The new fastest dual-GPU score submission in the 3DMark Fire Strike Extreme benchmark is 32,175 marks while in 3DMark Fire Strike benchmark it stands at 49,402 marks.
You can check out the submissions for yourself in the links above and also take a look at the Rauf profile page to see what else this guy has been up to. His recent work means that he now has the honor of possessing 75 cups and medals and moves up to 7th place in the HWBOT league rankings with a total points haul of 2,351. Congrats!
Just a few days ago we saw the end of the penultimate Rookie Rumble contest of 2017. Rookie Rumble #50 spanned the months of November and December and featured 482 Rookie HWBOT members, all vying for a crown that once again belongs to Skylead from France. He makes his mark once again, taking his third consecutive Rumble win, plus his first AMD rumble win. Nice work sir! It’s not often that we see one overclocker take wins in both contests. Let’s take a look at each of the stages, the hardware being pushed and the scores being produced in a little detail.
Rookie Rumble #50: November 18th - December 9th, 2017 - First, a quick reminder about what the Rookie Rumble series all about. The central idea is to give Rookie, Novice and Apprentice-class HWBOT members a place where they can compete against each other on a level playing field. For this reason Enthusiast, Extreme and Elite Overclockers are not eligible to compete. Round #50 of the contest was set up with three stages featuring the following benchmarks; Intel XTU, the classic SuperPi 1M and the GPUPI for CPU 1B benchmark. Let’s take a peek at each stage in a little more detail, starting with XTU.
Stage 1: Intel XTU - Interestingly we find that this round’s champion is not to be found in the top three of the Stage 1 leaderboard. Instead we find Skylead (France) in fourth place with a score of 477.5 marks per core (1,910 total) while klause (Germany) takes top spot with a score of 484.83 marks per core (2,909) using a Core i7 8700K ‘Coffee Lake’ processor which he pushed to a very impressive 5,260MHz (+42.16%). Not bad going at all. His motherboard of choice was the ASUS ROG Maximus X Hero (WI-FI AC) which was joined by a 32GB G.SKILL Trident Z RGB CAS 15 DDR4 kit.
Read the full roundup of the Rookie Rumble #50 contest here on OC-ESPORTS.
Last weekend saw the very last stop of the HWBOT World Tour 2017, landing in Berlin, Germany for the Overclocking World Championship Final 2017. The Final invited nine of the world’s best extreme overclockers to compete for cash prizes and the right to be World Champion 2017. Today we bring you the entire photo album of the event which if nothing else, reveals the fact that everybody had a really good time. Thanks to OverClocking-TV for the pics. Enjoy!
HWBOT OC World Championship Final: December 9th-10th, 2017
The HWBOT World Tour 2017 visited ten countries around the world this year. At each stop an Overclocking World Championship Qualifier contest was held, an extreme overclocking contest where the region’s most talented overclockers went head to head to compete for a seat in the Final. Here are some shots of the guys as they are prepped by HWBOT contest organizers and generally getting to know each other, plus shots of the trophies that were lined up for the winners.
Last week saw the very last stop of the HWBOT World Tour
You can find the full photo album with some very interesting shots of the happenings at CaseKing HQ in Berlin here on the HWBOT World Tour site. You can also find our full report of the contest here in Part 1, and Part 2 of the OCWC 2017 Final roundup series.
Last week we covered a video from Roman ‘der8auer’ Hartung where he started his mission to actually get to the die of a $2,000 USD Intel Core i9 7980XE processor. Don’t worry, he wasn’t attacking a new and functional chip. It was actually a broken one that someone passed on to him with the sole purpose of destroying it in order to retrieve the actual die. He returned yesterday with Part II of the video.
So just to recap. First he delidded the CPU using a specialized Skylake-X version of the Delid-Die Mate tool. He then cleaned up the chip, removing the liquid metal with acetone. Going outside to avoid breathing moxious fumes, Roman then heated the chip up to temperatures beyond 450 degrees Celsius. This allowed him to actually remove the die from the silicon package, an job that proved to be rather difficult due to the underfill layer of the chip which will not budge unless you get it really hot. Eventually Roman plans to do an examination of the die itself using a USB microscope which is capable of 220x zooming and image polarization.
Before that however he wants to remove the copper layer that remains on the chip. To this submerges the chip in 40% iron(III) chloride (or ferric chloride if you prefer) using an ultra sonic cleaner with temperature control functionality. This ensures that the iron (III) chloride is in the correct temperature range. After around two hours, you can see how the copper layer is starting to wear away, revealing some the die structure beneath.
What you end up with after all this work, is the actual die of the CPU can be seen, although you have to zoom in and look at it in the right light. The next step is to use glass etching paste, a similar process to that used above, just more a bit more dangerous. Not surprising as it contains hydrofluoric acid. After ten minutes or so, the CPU die is fully recovered, and looking pretty cool (as in the photo above). It’s fascinating to see the various parts of the CPU actually revealed on the die.
You can watch the video for yourself, here on the der8aeur YouTube Channel.
It must have been around 2013 when smartphones, particularly ones using Google’s Android OS, really came of age, becoming almost totally ubiquitous. It is also around this time that it became relatively simple to adjust the settings of your handheld device to make it go a little faster, and of course there were benchmarks around where you could test and compare your performance. Yep… I refer to the advent of Mobile Overclocking, the topic of a post written in December of 2013 by our very own Pieter-Jan 'Massman' Plaisier. Here are just some of the thoughts expressed. Thoughts which ultimately led to the development of the Mobile HWBOT Prime benchmark and the integration of Mobile Overclocking with HWBOT.
The idea is simple. As a community, overclockers have been able to force hardware manufacturers to care about the product quality. Through overclocking – how irrelevant the benchmark scores may be – and the competitive nature of the overclockers, we motivated marketing teams primarily, and engineering teams secondarily, to look at how to improve the design of their product. The companies wanted not only to prevent the power users from spreading the word on poor design, but also to win the race to feature in the world record system. The result we know: better bios, better hardware, more tuning, and better design. A win for everyone!
This eco-system does not exist for mobile devices. There are tons of applications for mobile architectures outside the space of smartphones and tablets to be uncovered. We cannot let poor hardware design stop us. Let’s kick-start the eco-system! The proposed trajectory is as follows. First we introduce the competitive spirit through a benchmark application. The open-source Android version of HWBOT Prime seems to be a good start. The hope is that through rankings and leader boards, developers get interested. Who can build the fastest ROM? Who can build the most overclockable kernel? We hope that in the Hackerspace we can find a couple of people who can help work on a specific device project. For us, it will be one of the Hardkernel Odroid devices. Mainly because we have a bunch of them, and they are easy/easier to work with that smartphones or tablets. Especially when it comes to experimenting with different types of cooling.
You can read the full piece from December 10th 2013 here, which also talks about the Taipei Hackerspace.
Having covered Part 1 of the Overclocking World Championship Final 2017 contest, we now turn our attention to Part 2 and the Elimination Phase of the contest.
Day 2: 1v1 Elimination Phase - The second phase of the contest is a little more complex than usual as it uses an elimination, nine player format that ultimately means it is possible to lose a few 1v1 matches and still go on win the contest. The rankings from the previous day dictate when each contestant will participate and how many matches they will eventually have to compete in. 8th and 9th ranked players from the Qualification Phase start first, meaning they may in theory have to win more 1v1 matches than the other contenders to make it through to the decisive and final Match 16.
To make things a little more interesting, we’ll cover the contest from the experience and perspective of each individual overclocker, taking a look at the matches they competed in and the resulting outcomes. Remember, to better understand the flow of the contest, you can refer to the completed brackets by scrolling down to the bottom of this page. Let’s start with jordan.hyde99, Australia’s one big hope!.
jordan.hyde99 (Australia) - Jordan arrived at the contest as arguably the undisputed newcomer to Elite level competitive overclocking. For more info about Jordan, read the jordan.hyde99 bio in this profile article we did in the leadup to the contest. His performance in the Qualification Phase placed him at the foot of the table, meaning he faced PXHX in the first match of the contest in Round 1. This ended in a loss as he failed to make a valid score in the XH265 4K benchmark, while the Brazilian managed a score of 14fps. Perhaps the fairly long benchmark run involved with the 4K preset presented a problem for Jordan. Time management might well have been an issue here.
Read Part 2 of the OCWC 2017 Final roundup article in fullhere on OC-ESPORTS.
Just under a week ago NVIDIA launched its new Volta architecture graphics cards series, surprising a few onlookers by launching a $3,000 USD GeForce GTX Titan V that uses a ‘Big Volta’ GV100 GPU. So far a few HWBOT members have stumped up the asking price and started benching with the new latest and greatest from NVIDIA. Let’s take a look at handful of score submissions which include one World Record and 4 Global First Place, single card scores.
H2o vs. Ln2 (US) Breaks Catzilla 720p World Record - The original Mr Slinky, H2o vs. Ln2 from the US has acquired at least a few cards and has already broken his first World Record. The new VRMark Cyan Room World Record stands at 14,806 marks and was made with an NVIDIA Titan V at (apparently) stock settings, i.e. the GPU with a boost frequency of 1,455MHz and graphics memory at 850MHz. The rig also used an Intel Core i9 7980XE 'Skylake-X' chip, which puts the price of the rig's two key components at a tidy $5,000 USD alone.
H2o vs. Ln2 has clearly also been experimenting with his monster liquid cooled rig, pushing both GPU and CPU to break a few Global First Place scores. In the Unigine Superposition - 1080P Xtreme benchmark, the new fastest ever score with a single GPU now stands at 10,446 points. This was made with the Volta-architecture GPU pushed (according to the submission post) to 2,010MHz with memory at 2,132MHz. The CPU was also apparently pushed to apparently to 5.5GHz. The same rig also broke the Global First Place score for a single card in GPUPI - 1B, which now stands at 4sec 65ms.
jpmboy (US) Breaks 3DMark Time Spy and Time Spy Extreme GFP Scores - US Overclocker jpmboy has also been enjoying some Titan V performance breaking the 3DMark Time Spy Global First place record for a single card with a score of 15,570 marks. He pushed his GV100 GPU to 1,546MHz with graphics memory at 1,455MHz (+71.18%), also using an Intel Core i9 7980XE 'Skylake-X' pushed in this case to 4,900MHz. In the 3DMark Time Spy Extreme benchmark the new GFP score for a single card now stands at 8,180 marks.
NVIDIA’s new Volta GPUs certainly appear to offer a new level of performance compared to Pascal. It will be interesting to see how the GeForce GTX 2080 card performs when (if?) it arrives. Until then you can find all the new Volta scores in the links above. Enjoy.