AMD R9 Series Logo

One of our most reliable sources, SweClockers,com has reported that the AMD Tonga GPU will actually be poised to replace the R9 280 and R9 280X respectively. That means that when we learned of the existence of this class a few days back thanks to our friends over at the exact tier was speculated to be mid-range at best. However this new information seems to suggest that Tonga is actually the Early-High End Lineup of AMD.

AMD Tonga GPUNot an Official AMD Slide.

AMD Tonga GPU could have 265 Bit Bus Width and 7Ghz vRam – Potential Nomenclature is Radeon R9 285

So basically, AMD Tonga GPU will replace the now aged Tahiti core, it also means that performance wise, it will be faster than the current Maxwell cards; namely the Geforce GTX 750 and GTX 750 Ti. According to the same source, the circuit actually looks like a scaled down variant of Hawaii albeit with a 256 Bit memory bus. AMD will probably compensate for this by using faster Memory Modules so 7 GHZ GDDR5 RAM is definitely on the table. We also no have word on pricing though the expected launch date is third quarter of 2014, probably August or September.


Word on the street is that Galaxy will soon make a watercooled addition to it’s Hall of Fame premium card series. The HOF series was launched by Galaxy in 2010. It consists of cherry picked reference GPUs with design upgrades to achieve greater performance specifications. According to Galaxy, the cards have been solely designed to enter the 3D Mark Hall Of Fame. HOF cards are generally air-cooled and this is what makes the appearance of this waterblock interesting. It now most likely means that a watercooled HOF will be released somewhere in the near future.


Waterblock in Traditional White Galaxy HOF Design. Can Cover Entire PCB.

Courtesy of our friends over at, we have been able to get our hands on some pictures of the waterblock. It is rather elegantly designed, with the main portion of the waterbody being white. Only the metal plate with the model name is in metal. The color choice of the card falls in line with the HOF. However, that is the only thing that it sticks to. This reveal marks the first watercooled GPU of the HOF series. The body, as we can see in the picture above is quite large. This will allow the waterblock to be fully covered, in addition to efficient cooling of the GPU. Also, this complete covering of the PCB also ensures that the graphic card will remain stable during performance intense operations. This is because the memory and power supply components will also be covered and give more room for overclocking.

AMD Radeon logo

This week, Russian enthusiasts Smoke and Atheros have managed to overclock AMD’s Radeon R9 290X. The core frequency on the card was pushed to 1575 MHz. The R9 290X has a stock frequency of 1000 Mhz and is based on Hawaii. It is coupled with 2816 stream processors and also has 4 GB of GDDR5.


R9 290X Core And Boost Frequency Overclocked to 1500/1700 MHz Respectively. First Position in 3D Mark 11 with 23,260 Points

The R9 290X has stock core frequency specification of 1000 Mhz and a turbo frequency of 1500 MHz. Smoke and Atheros have managed to push this to 1500 and 1700 MHz with a power rating of 1.475/1.7 V. SLI/Crossfire on the card was off. This makes for the first position in the 3D Mark 11 benchmark with a score of 23,260 points. It also makes for a Global Rank of 6 for 1x GPUs. The number of UHP points gained using R9 290X is 18, which also gets a top global position. The processor used in the overclocking process is Intel’s Core i7-4960x. The processor has been officially rated for a frequency of 4 Ghz (3600 MHz). It has 6 cores with LGA2011 and a TDP rating of 130 W. The 3600 core frequency on the processor was overclocked to 5893 MHz. Through liquid nitrogen, the temperature was kept at -110 degrees.


You would be forgiven for thinking that this particular Black Edition means 6GB worth of GDDR5 memory because of the similarity to Titan Black, but this particular GTX 750 Ti Black Edition is just a very market aware use of nomenclature on Gigabyte’s part to sell a particularly good variant of the Maxwell GTX 750 Ti. Of course the Wind force 2x Cooling system will leave a lot of headroom for over clocking as well, not to mention the rigorous stress testing it has gone through means some serous stability for server uses.


Gigabyte Launches GTX 750 Ti BLACK Edition with 2GB GDDR5 Memory and Windforce 2X Cooling System – Targeting Server Market

The card has no relation to the GTX Titan Black in any way whatsoever. However these series of cards had to pass 168 Hours of difficult stress testing to earn the “black” suffix and will come with a certificate of the same. This will mean that this hardware is server ready because a 168 Hour long stress test is usually sufficient for even the most difficult of circumstances. This card is very subtly aimed at the server market as well as the usual game market. The Maxwell 750 Ti is one of the two most energy efficient GPUs in this spectrum and with the high stability rating, it would be a good option for server grade work. For example these would be ideal in a low level server or a cluster compute setup using Maxwell 750 Ti Blacks and LGA 2011 Intel parts. Cost effective while keeping the power consumption under check.

NVIDIA GeForce GTX Titan Z

NVIDIA has got in some deep mess with their GeForce GTX Titan Z as the launch of the card was planned for April 2014 yet the card hasn’t been launched in market as of yet. If that’s not enough, a review of the card leaked just a few days ago which showed lower performance than its much lower priced competitor, the Radeon R9 295X2.NVIDIA GeForce GTX Titan Z Retail

Image Courtesy of Overclockers.Co.Uk!

Retailers Await NVIDIA’s Greenlight For GeForce GTX Titan Z Launch

Now you must be wondering what went wrong with the launch? Performance and design aside, it has been said on several occasions that NVIDIA delayed the card due to unfinished drivers since they want their card to be out with proper support for games and applications. The development of software for the NVIDIA GeForce GTX Titan Z has led the card to face several delays and it still remains unknown when does NVIDIA plan to launch the card officially.

While there has been some misinformation spread on several forums and communities that the card has been cancelled due to poor performance, that’s not true at all. The card is ready for deployment and several retailers have access to the GeForce GTX Titan Z in their warehouses. reports that they have already got retail units of the card ready for launch and are awaiting NVIDIA’s greenlight to officially start selling these cards. Gibbo over at the OCUK Forums has the following to say:

Tegra K1 Logo

NVIDIA revealed their Tegra K1 SOC at CES 2014 which will feature the latest Kepler cores along with the 32-bit ARM and 64-bit Denver cores. Yesterday, Xiaomi unveiled their upcoming Xiaomi MiPad that is a 7.9″ tablet and the first device to feature NVIDIA’s Tegra K1 system on chip.Xiaomi MiPad Tegra K1

Xiaomi MiPad To Ship With 32-Bit and 64-Bit Denver Powered Chips

Before heading to the main news, I would like to talk about the Xiaomi MiPad which was introduced yesterday. The MiPad which is a 7.9″ tablet features an IPS display bearing a resolution of 2048×1536 and surrounded by a asymmetrical bezel, plus being available in several different colors which are a nice touch to its design scheme. The design specifications include an 8MP main camera on the back followed by a 5MP camera on the front and the design is mostly simplistic.

On the technical side, the Xiaomi MiPad is the first tablet to feature NVIDIA’s Tegra K1 SOC and would be sold in both 32-bit and 64-bit variants. The Tegra K1 SOC is built on a 28nm HPm process which holds true for the both variants. The dual core variant with Denver CPU and 7 Way superscalar compute will feature the 64-bit ARM v8 cores and 192 Kepler cores with clock frequency of 2.5 GHz while the 32-bit Quad core ARM variant will be clocked at 2.3 GHz. The 64-bit Denver model comes with 128KB + 64KB while the 32-bit variant comes with 128KB + 32KB L1 Cache. Both models will feature a 25.46 Whr (6700mAh) battery, 16 – 64 GB flash storage, Micro SD slots, 802.11 b/g/n wireless and Bluetooth 4.0. Both devices would feature support for Android 4.4 KitKat. Arriving in June 2014 (just around Computex), the price would be set around $240 for the 16 GB and $272 for the 64 GB variant.

AMD R9 Series Logo

AMD might possibly be preparing their upcoming Tonga GPU to rival against NVIDIA’s GM107 chip based on the Maxwell core, as reported by Videocardz. Just a few days ago, we heard about AMD’s upcoming Volcanic Islands 2.0 graphics cards which include Tonga, Maui and Iceland along with AMD plans to feature HBM High Bandwidth on some of them.AMD Tonga GPU

AMD Possibly Preparing Tonga GPU Architecture For Radeon Cards

The codenames of Tonga, Iceland and Maui have appeared in Catalyst files several times but their reemergence means that AMD has had these chips under development phase for quite some time which may explain why we didn’t saw any new GPU architecture in the sub $299 range of the Radeon R200 series cards. The source reports that the Tonga GPU will not be based on a 20nm architecture nor would it feature the high-bandwidth memory which is due to several reasons.

The Tonga GPU architecture is being built on a new architecture which is presumably the GCN 2.0 design and not the GCN 1.1 architecture which has been featured on Bonaire and Hawaii. The Tonga graphics architecture will be built on the 28nm node from Global Foundries and would feature a new design scheme introducing latest architectural improvements such as new ACE (Asynchronous Compute Engine) Units and more focus towards compute shaders. It is also mentioned that Tonga will retain the basic technologies of the Radeon lineup such as Mantle, TrueAudio and XDMA for CrossFire support.


AMD has announced today a price cut on it’s recently released Radeon R9 280. The R9 280 was announced this March and is one of the company’s lower performing GPUs. Compared to the next R9 280X, the next GPU in the lineup, there are several performance differences between the two. The 280 has a 20% lower clock speed of 5000 when compared to the 280X. In addition, the GPU has a lower turbo clock speed of 925 MHz as well. Other performance differences include lesser shading and computing units and also a lower memory clock speed.


Cryptocoin Mania Subsides. The R9 280 Back to Price Relative to Performance Specification of Other Cards.

The 280 was launched by AMD in March. This was amidst all the bitcoin – litecoin mining craze, In February particularly we saw the 290X’s price briefly touch the $ 900 mark. Considering it’s MSRP was $ 550, this is a substantial increase. AMD’s GPUs were known to perform better at mining, giving more hashing speeds when compared to rival Nvidia’s GPUs, with the R9 290X giving a speed of 790 MH/s. Since all the hype is over now, AMD is now cutting the price of the card to $ 249. It should be kept in mind that amidst the price rocketing of the cards, AMD had not changed it’s MSRPs. This not only led to the company not getting any of that additional dollar change in the price but also had other affects. AMDs entire strategy had evolved around GPU price cuts to increase sales. Some of its GPUs equaled Nvidia’s in performance but were available at lower prices. This increase in overall prices did indeed put the company in quite a pickle.

NVIDIA GeForce GTX Titan Z

NVIDIA GeForce GTX Titan Z review has leaked prior to its launch sometime in Q2 2014 and things don’t look well for the upcoming flagship card from NVIDIA. The GeForce GTX Titan Z is the first graphics card solution featuring NVIDIA high-performance GK110 core architecture which has spanned an entire generation as the GeForce 700 series high-end cards namely the GeForce GTX 780, GeForce GTX 780 Ti, GeForce GTX Titan, GeForce GTX Titan Black and now the GeForce GTX Titan Z. The benchmark was leaked by a Hong Kong based tech site which has been revealed courtesy of Videocardz.NVIDIA GeForce GTX Titan Z

NVIDIA GeForce GTX Titan Z Review Leaks Prior To Launch

The shocking thing is that there’s no conformation about the cards launch with the reason blamed to be the final drivers which we were are starting to believe would be false and has more to do with the performance of NVIDIA GeForce GTX Titan Z which lacks severely behind its competitor such as the AMD Radeon R9 295X2. The card has been delayed two times during 29th April and 8th May. It looks like the earlier rumor about the boost speeds are true and the card is clocked at 1058 MHz but let’s take a brief look at the specifications before going into the performance section.

The GeForce GTX Titan Z is manufacture red with two GK110-350-B1 chips under its hood that pack 7.1 billion transistors each. The GeForce GTX Titan Z replaces the GeForce GTX 690 boasting dual-GK110 cores compared to dual-GK104 cores on its predecessor. The GeForce GTX Titan Z will feature two GK110 cores with 5760 Cuda Cores, 448 TMUs and 96 ROPs. The card features a 384-bit x 2 bus which will run across a massive 12 GB VRAM. This is an impressive feature giving developers and games an unprecedented amount of VRAM for use. The memory is clocked at 7 GHz effective clock speed. The core clock speeds are maintained at 705 MHz base and 876 MHz boost clock and the card features a maximum single precision performance of 8.1 TFlops and 2.3 TFlops of double precision.

AMD R9 Series Logo

While NVIDIA plans to launch their high-end Maxwell architecture in the second half of 2014, AMD has also plans to introduce their Volcanic Islands 2.0 graphics cards during the same time. Forums have got some interesting details to share regarding the upcoming generation of AMD GPUs and from the looks of it, AMD could be the first graphics company to adopt the HBM Memory design.Die stacking and system

AMD Could Feature Next Generation HBM Memory on Volcanic Islands 2.0

According to a slide revealed by the forums, AMD has already started sampling first generation HBM with their graphics cards. The specifications of the HBM memory are interesting featuring a voltage of 1.2 – 2.5V with the DRAM Die density of 2 Gb per stack, each stack featuring 4 DRAM modules. The bus interface would be 1024-bit wide and the command interface would be the traditional DDR (GDDR for graphics units). Each stack would be made up of a logic die which will feature a 2.5D or 3D interface on which four DRAMs would be stacked. This stacked layer would be fused on the PCB just like the regular memory chips but would deliver high performance as the name suggests (HBM = High Bandwidth Memory). The slide is from December 9th 2013 and is called “Die Stack is Happening” which might give an idea why we have seen so many rebrands within the initial Volcanic Islands launch lineup till now. The following chart shows some difference between GDDR5 and HBM Stacked memory designs:


Disclaimer: The Conclusion given at the end has been given in a completely non-professional capacity. Though the Guidance has been given to the best of my abilities we cannot be held liable in any way whatsoever. Nvidia’s (NASDAQ: NVDA) first quarter fiscal results are out and they have reported an EPS of $0.29 they have successfully beaten analysts expectations of $0.17 (+$0.12) cents which is approximately double. The PC Market is under continuous allegations of being on a generally downward trend but despite of the bearish market, Green managed to pull through. However, the announcement saw some shorts pulling out in an early stock dip.

Nvidia Corporation (NVDA) Q1 FY15 Financial Results are Out – Beats Expectation by 12 Cents, General Growth Despite Bearish PC Market attributed to GPU Sales

NVIDIA Corporation 5th May to 8th May Stock PriceNvidia (NASDAQ: NVDA) has seen some impressive sales figures in all its departments this quarter but none of them come close to the sales of the GPUs which account for 81% of the entire Revenue this quarter. However even though GTX sales saw an increase of 14% Y/Y they went down 4% sequentially. On the other hand the Tegra SoC Division accounted for 12.6% of revenue and grew 6% sequentially and 35% yearly, which is very decent growth. The reason for the same is attributed to the increased demand from the automotive industry (+60% Y/Y) and smartphone Industry (partial decline). GRID Computing and Tesla Sales remained strong due to VDI Deployments and an assortment of streaming gaming providers. However the fact remains that both the Tegra and GPU Division accounted for  93.6% of the entire revenue of the business. The increased demand for Maxwell (28m) i.e. the GTX 750 and GTX 750 Ti is attributed to the sales while the notebook sector saw a slight decline in demand. Here is the financial summary of the quarterly:

NVIDIA GeForce GTX Titan Z

Yesterday, NVIDIA was supposed to launch their high-performance GeForce GTX Titan Z graphics card however the card isn’t available in the market nor could the reviews be found online. This is a bit disappointing and delay has been confirmed by GDM.OR.JP who mention that the GeForce GTX Titan Z has been delayed indefinitely and no words on the new launch date have been passed over to the press.NVIDIA GeForce GTX Titan Z

NVIDIA Delays GeForce GTX Titan Z Launch Once Again

NVIDIA’s GeForce GTX Titan Z which is composed of two GK110 cores was initially planned for launch of 29th of April however due to unfinished drivers and a design change to the bracket, the card was delayed to 8th April 2014. The more recent reports are now saying that retailers in APAC market have mentioned that the card has once again been delayed to an unspecified date and the reason is once again blamed to be the drivers. It is not know what driver would feature but it probably boosts the card and optimized specifically for the GeForce GTX Titan Z. Here’s a small description of the driver from our previous post:

The NVIDIA GeForce GTX Titan Z is planned for launch on 8th May 2014. The reason for the delay has to blamed on the NVIDIA drivers which were not ready to be shipped during the last month. The new official drivers which are named GeForce 337.69 is a press driver and a WHQL variant would be available to consumers during the launch day.

richland apu a10-6790k

A few days ago we revealed the entire Mobile Kaveri and Crystal GPU Series and I promised that I would be revealing benchmarks within 24 hours, but due to a slight technical malfunction (my laptop threw a fit) the actual delay was slightly longer. So sorry about that, but anyways here are both the Gaming and Compute benchmarks for AMD’s newly revealed APUs. And I have also thrown in the APU Generation Transition Table as consolation for the delay. Without further ado, lets dig in.AMD Mobile Kaveri APUs Gaming Performance

ULV Mobile Haswell Pit Against Mobile Kaveri – Results Show Very Promising Results for AMD’s Low End Spectrum

The AMD Mobile Kaveri APUs were pit against the ULV (Ultra Low Voltage) Spectrum of Intel Mobile Haswell CPUs.  While we did not receive the exact specifications of the testing rig, rest assured that every bottleneck was neutralized. And since I have delayed this enough, I will skip the preliminaries this time.
Now these are very interesting benchmarks. The OEM-only flagship FX-7500P was not benchmarked, however because the Retail-Available FX-7500 is ahead of the ULV Haswell by leaps and bound we can very safely assume that it will be even further ahead. Not only that but the very-much mainstream A10-7300 APU is comfortably ahead of the i7-4500U, i5-4200U and i3-4010U in terms of Gaming Performance. What it actually means is that APUs are finally exiting the territory of HTPC-only setups and entering the realm of true gaming power. While you would not be able to run any modern game at anywhere near the modern games, there should be quite a few games out there that run well at low-medium settings and considering the low power consumption. That is definitely a bargain.


Our colleagues over at have spotted some very interesting information on the Chinese forum Chiphell. Apparently an import/export data sheet suggests that the GTX 880 might actually have 8GB of GDDR5 RAM. Now it goes without saying that this information at the moment is not anywhere near concrete news, so take it with a pinch of salt.

NVIDIA Maxwell

Nvidia Geforce GTX 880 Could Actually Ship with 8GB of GDDR5 Memory – Water Cooling spotted as well

I have a very healthy wariness where speculation is concerned, but this comes from reliable sources so it is definitely worth posting about. Basically the specs sheet of the import/export data from Zauba is available there and something very interesting is present in those sheets. One thing that VCZ noticed was that the memory mentioned, is in fact 8GB GDDR5 and another thing that I noticed was that the components look like a DIY package of sorts. I.e. anything int hat list could very well end up on the GPUs. And the spec sheet lists a “water cooled plunger” too and if that doesn’t raise eyebrows I don’t know what does. Now according to the same spec sheet, the GM204-B1 will power the GTX 900 series, which will be Maxwell on 20nm. Similarly the GM204-A1 and GM204-A2 are apparently reserved for the GTX 800 Series, which might or might not be at the 20nm node (unlikely).

PowerColor logo

A few days ago, PowerColor teased their upcoming Radeon R9 295X2 Devil13 which would be the successor to the previous Devil 13 card that featured two Tahiti cores. The card will probably be the first custom cooled R9 295X2 based card since we have only seen reference variants of AMD’s flagship dual-chip card in the market.

Something very special coming up in May… dare to guess? — PowerColor

PowerColor Radeon R9 295X2 Devil 13 Power ConnectorsPowerColor Radeon R9 295X2 Devil13 Unveiled – Requires Four 8-Pin Connectors

Now I know this may sound crazy but Chiphell Forums actually managed to get pictures of the card prior to its launch and from what I am currently looking at, the card is a total beast in terms of design and performance. PowerColor’s Radeon R9 295X2 Devil 13 will feature a custom designed PCB which we can’t see at the moment but its really beefy in terms of design. Before moving into the design, we will do a round-up of the specifications.

The PowerColor Radeon R9 295X2 Devil 13 which is also codenamed ‘Vesuvius’ (Official AMD naming) features two Hawaii XT cores with GCN 1.1 architecture. With two Hawaii XT GPUs on the same board, the specifications equate to 2816 x 2 stream processors, 176 x 2 texture mapping units, 64 x 2 Raster Operators and 4 GB of GDDR5 memory on each chip which equates to a total of 8 GB GDDR5 memory. Double these and you get an impressive 5632 stream processors, 352 TMUs and 128 ROPs for an unprecedented amount of gaming and compute performance. The 8 GB GDDR5 memory runs across a 512-Bit X 2 memory bus that is clocked at 1250 MHz effective clock speed (5.00 GHz effective clock speed). This pumps out a total of 640 GB/s memory bandwidth. Core clocks are not mentioned but i suspect an insane factory overclock which will still allow users to crank some high numbers on the card.

Creative Commons License This work by WCCF (Pvt) Ltd. is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.
Permissions beyond the scope of this license may be available on the Terms and Conditions page. WCCF 2011 is designed by UzEE, Inc.
Ran 21 queries in 1.131  seconds.