Nvidia GeForce GTS 650 is Only 3% Faster than GTS 550 Ti

Nvidia’s new Kepler architecture has proven to be quite a success among gamers and this is owed to the fact that the GPU is very efficient, relatively cool and overclockable, next to the impressive performance it brings.

During the initial launch, many mid- and low-end cards based on the new architecture were announced, but not so many managed to reach users, as Nvidia is first waiting to clear the old Fermi-based stock. Slowly but surely, the Kepler GPUs are replacing the Fermi architecture in the company’s lineup and now GeForce GTS 550 Ti’s replacement is close to launch, as VideoCardz report. This card that will replace the popular GTS 550 Ti is called GeForce GTS 650 and is powered by the GK107 Kepler GPU. Compared with the GTS 640 we already know, Nvidia’s GTS 650 comes with the same number of 384 CUDA cores and 128-bit memory interface, but it will likely be clocked at 1 GHz or more.

The memory is GDDR 5 with a likely effective data rate of 4800 GHz instead of the DDR3 used on the GT 640. The maximum power consumption is rated at about 70 watts, which means that reference cards will not feature the extra power connector that GTS 550 Ti has. Compared with the 192 pipelines that the GF116-based GTS 550 Ti has, Nvidia’s GTS 650 will only be about 3% faster while consuming considerably less energy than the 116 watts the GF116 is sucking. There is not much info on the benchmark result, but the guys that managed to test the card confirm that the performance difference is almost nonexistent.

Nvidia Low-end comparison
Images credits to vga.it168

Read more »

Team Group UHS-1 Xtreem Series Memory Cards

We don't know when, where and for how much money you'll be able to buy one of Team Group’s newest memory cards, but we do know everything else worth knowing.

Called SDHC UHS-1/Micro SDXC UHS-1, they are made for tablets, phones and driving recorders. As suggested by the UHS-1 specification, the transfer speeds are significant: up to 40 MB/s reading, 10 MB/s writing. Speaking of which, random read IOPS are of up to 500, while writing is done at 150.

Furthermore, the data retention is of over 10 years and the power source of 3.3v. All in all, the newcomers are about three times better at writing than products of the same class, or at least three times faster at scribing bits. Real-time communication apps and boot loading of portable gadgets should exhibit the sharpest time differences.

Team Group UHS-1 memory cards
Image credits to TechPowerUp

Read more »

ADATA XPG Gaming v2.0 Series DDR3 2400G Memory

ADATA is probably in a good mood after its results of last month, and there aren't many better ways for such an IT corporation to celebrate a job well done than a new and mighty piece of hardware.

We say ADATA is likely in good spirits due to recent findings that suggest it has good product shipments and high revenues in June 2012. The company is working hard to ensure the third quarter is at least as productive, although this newest product might not help as much as one would think. We aren't implying that there is anything wrong with it, far from it, but high-end hardware tends to not reel in many users, due to its price. Still, the XPG Gaming v2.0 Series DDR3 2400G DRAM modules will serve just fine as enhancers of the company's prestige.

For those who don't know, XPG stands for Xtreme Performance Gear. It is a brand with a similar role to ASUS' ROG. The maximum frequency is, obviously, 2400 MHz / Mbps, while the bandwidth can go as high as 19200 MB/s. Also, the modules, with their capacities of 4 GB, have XMP certification (Intel Extreme Memory Profile) and the Thermal Conductive Technology (TCT). Whatever heat isn't dissipated by the 2oz copper 8-layer printed circuit board, the heatspreaders take care of (they have a screw-lock mechanism that boosts cooling efficiency for long-term use).

Finally, ADATA managed to squeeze a latency of CL10-12-12-31. Latencies are always less impressive on modules with high frequencies but this setting is actually more than decent. The product page of the ADATA XPG Gaming v2.0 Series DDR3 2400G DRAM modules is right here. If you live in the US or Canada and own a Z77 motherboard and a third-generation CPU, you might want to consider paying the $89.99 / €73.23 price.

ADATA XPG Gaming v2.0 Series DDR3 2400G
Images credits to ADATA

Read more »

CyanogenMod 10 Preview Builds Available for Samsung Galaxy S III

As soon as Google announced the availability of Android 4.1 Jelly Bean, the CyanogenMod team promised the availability of CM10 based on the new platform release.

CM is one of the most popular custom software for Android devices out there, and many are already eagerly waiting to get it on their units. For some Galaxy S III users, the wait is over, though with a caveat. Only preview builds of CM10 can be downloaded at the moment, and they are buggy.

However, things have started to move in the right direction, and more stable builds are expected to become available in the not too distant future. In the meantime, those who would like to grab the early software release should head over to this thread on XDA-Developers for the SGH-I747 (LTE) downloads, and to this thread for the T999 bits.

Samsung Galaxy S III
Image credits to Samsung

Read more »

Here Are Some Intel Core and Pentium Series CPUs Set for September Release

If Intel isn't launching anything, the press and Internet will look into what products it will soon or eventually produce, such as Core and Pentium series CPUs.

Last week, for example, we saw how Intel's Core i7-3630QM and i7-3632QM were set for Q4 launch. Now, we are going to detail seven more central processing units, five of which belong to the Core i3 series and two of which are Pentium chips. All of them have four things in common: the thermal design power (TDP), the architecture (Ivy Bridge), the number of cores (two) and the cache memory (3MB). Everything else is different enough to justify the distinct product and series names. The Pentium G2100T 2 and Pentium G2120 lack hyper-threading and run at 2.6 GHz and 3.1 GHz, respectively. The former has a TDP of 35W, but the latter needs 55W.

The five Core i3 CPUs are called Core i3-3220, i3-3220T, i3-3225, i3-3240 and i3-3240T. Only the Core i3-3225 has HD 4000 graphics, while the others make do with HD 2500. Since they possess Hyper Threading, the five processors have 4 threads each. That said, Core i3-3220 is a 3.3 GHz part at 55W TDP, while the Core i3-3220T sacrifices some performance (2.8 GHz) in order to reduce the energy requirements to 35W. The Core i3-3240 and Core i3-3240T have 3.4 GHz and 2.9 GHz speeds, respectively, and similar TDPs are their siblings. As for the Core i3-3225, it is a 55W part with a 3.3 GHz speed.

All of them will be formally launched in the second week of September 2012, which, since the month begins on a Saturday, means between the 2nd and 8th of the month. The folks at CPU World are the ones who uncovered and compiled most of this information, so you can go there to read about the products more at length.

Intel Core i3 logo
Image credits to Intel

Read more »

Galaxy Overclocked GeForce GTX 670 with 4GB Memory

In the fight to convince the user to pick a certain product from the store shelf instead of a competing brand, Chinese video card manufacturer Galaxy is preparing a factory overclocked version of Nvidia’s GTX 670 GPU.

The new card is called Galaxy GeForce GTX 670 GC Edition 4GB and just like its name says, it comes with 4 GB of GDDR 5 video memory. As Overlockers.ua report, The GPU frequency has been increased from Nvidia’s reference default of 915 Mhz with 1006 MHz boost to a 980 MHz with a 1084 MHz boost.

The memory frequency has unfortunately been left at the same 6008 MHz data rate level, but the user is free to try and overclock them further. The cooling system is comprised of two 92-mm fans blowing air through a cooler containing four copper heatpipes with a 6-mm diameter. The price is estimated to be around $518, which is about €424 for the European gamer.

Galaxy GeForce GTX 670 GC Edition 4GB
Images credits to Overclockers.ua

Read more »

Raspberry Pi Restrictions Lifted, True Mass Availability

The Raspberry Pi, that very small critter that somehow manages to browse the web and play 1080p video, finally has no supply problems.

Raspberry Pi is an ARM-based system the size of a credit card. It was envisioned as a platform for Linux programming, but it caught on with the public in such a way that the press has been keeping its eyes on it like a hawk. It helped that the device can turn any TV into a Smart TV, as long as the display has an HDMI input. That said, the product sold out on several occasions. Even we had to wait for many weeks before our order reached us. Huge pre-orders played a significant part in the lack of availability. About 350,000 models were requested, even with a limited number of Raspberry Pi per customer.

There should be more than enough in stock now though. The Raspberry Pi foundation now makes around 4,000 a day. “Customers worldwide can now order multiple quantities of the Raspberry Pi Model B board, along with the associated accessories, including SD cards pre-loaded with the latest Raspberry Pi operating system and Raspberry Pi cases for safer storage. Customers will be provided with a forecast future delivery date when placing their order, and these orders will be fulfilled after all orders placed before 16th July have been shipped,” said a spokesperson from RS Components. To place an order, all it takes is a visit to http://pi.rsdelivers.com, so if you're a student, an HTPC enthusiast or just a curious experimenter, you can let loose, assuming you don't want the Odroid-X instead. 

As a recap, Raspberry Pi has a 700 MHz ARM11 processor, a bunch of ports (mini HDMI, USB, LAN, etc.) and 255 MB RAM (not that much but, then again, this is small enough to hold in a palm, so it's still quite a feat). “Element14 are pleased to announce that we will now be taking volume orders for Raspberry Pi Model B, on an expected delivery lead-time of 4-6 weeks, as our order backlog improves and our production capacity continues to increase,“ said element14/Premier Farnell.

Raspberry Pi 
Image credits to Raspberry Pi Foundation

Read more »

Sapphire Radeon HD 7970 GHz Edition Toxic 6GB

Advanced Micro Devices, by pushing the Radeon HD 7970 to 1 GHz, managed to oust the NVIDIA GeForce GTX 680 from its top spot, performance-wise, but Sapphire took things further.

We'd say that the new board is one step ahead of the curve, if it weren't more accurate to replace step with mile-long leap. The name of the Sapphire Radeon HD 7970 GHz Edition Toxic 6GB really says it all: 6 GB GDDR5 VRAM are present (the exact amount is 6,144 MB), instead of “just” 3 GB. The product is a fairly peculiar invention, and not because of the hardware per se, but due to the circumstances of its appearance. Advanced Micro Devices actually took the world by surprise. The HD 7970 GHz edition came out of left field for most, and had to go through some last-minute changes.

Sapphire's beast, on the other hand, has been in the making ever since CES 2012, months ago. Either the OEM foresaw AMD's move or just figured a super-card would look good on the resume. At any rate, the Tahiti 28nm GPU is clocked at 1,050 MHz (1,100 MHz boost frequency) while the VRAM works at 1,500 MHz (6.0 GHz effective). Twenty-four memory chips account for the colossal MB number, while an 8-phase VRM powers them and everything else on the 12-layer PCB. As for the cooler, it has a vapor-chamber plate and a heatsink around it, plus four nickel-plated copper heatpipes (2 x 8mm and 2 x 6mm) and a couple of 90mm high airflow fans. Finally, an anodized aluminum backplate cools the chips on the reverse side of the adapter.

Overclockers will have a field day with Sapphire's Radeon HD 7970 GHz Edition Toxic 6GB, especially once they figure out the TriXX software (tweaks clocks, fan speeds, voltages), diagnostic LEDs and dual-BIOS (one is a failsafe). The price was not mentioned, unfortunately, but it will be high. The monster has 2,048 Graphics CoreNext stream processors, 128 TMUs and 32 ROPs after all. Shipments will begin shortly, in limited quantities.

Sapphire Radeon HD 7970 GHz Edition Toxic 6GB
Images credits to Sapphire

Read more »

Nvidia GeForce GTX 660 Coming Next Month

Nvidia’s Kepler architecture is a very efficient design that allows the flagship card to achieve results equal or even better than AMD’s Tahiti cards, despite the fact that AMD’s chip has half a billion transistors more and a 50% wider memory BUS.

Usually, the chip designer wants to use all its chips, regardless of whether some functional units suffer from errors. Such a chip has those particular units disabled and it is sold as a lower performing model. This happens because yields are not perfect and the fabless company cannot pay for the whole wafer and throw out half of the chips. When Nvidia’s Kepler came into being, chip yields were particularly good and the company actually did not have enough chips to fulfill the demand for a card like GeForce GT X 660. Usually, the chip designer waits and piles up a significant number of GPUs with disabled units and only when it considers it has enough to ensure fluent supply, it launches the card.

On the other hand, if yields are good and most of the GPUs do not have numerous errors and can be set up to function as a GTX 670 or GTX 680, the company has no real interest in selling the same die for less, as long as they’ve already paid a great amount of money for the wafer processing. Nvidia’s particular situation is that they also have a successful line of GTX 570 video cards and they don’t want to kill sales or slow down the stock clearing of their GTX 570 GPUs. The GTX 660 is likely to feature a 192-bit memory BUS and just 1152 or even 1344 CUDA cores. The video memory is said to be set at the 1.5 GB level and the pricing is slated at about $300.

Nvidia Kepler Marketing Shot
Image credits to Nvidia

Read more »