Around the web it was a pretty standard week at PC hardware review sites. We’ve got a couple of GPU reviews, some odd balls, and even a CPU review in store for you this week. Who’s looking forward to SC11 conference?
Brent Justice of [H]ard|OCP got his hands on a pair of ASUS’s new ROG MARS II Limited Edition video cards this week. This particular card is a custom version of the GTX 590. It has a fully custom PCB design as well as a dual 120mm fan cooling solution, but what really separates this card from the GTX 590 is its core and memory clocks, which are slightly above the GTX 580’s reference clocks, rather than below the GTX 570’s, like the original GTX 590. Needless to say a pair of these cards in SLI is the fastest graphics solution on the planet, though it’s worth noting that that solution consumes about 1200 Watts under load. ASUS has made another super high-end graphics card, is anyone else looking forward to the ARES II?
SemiMD’s Mark LaPedus decided to analyze the DRAM server roadmap a bit, and found that things look rather muddled. Vendors are preparing for the DDR4 transition and beginning to implement the first 3D technology, through-silicon-vias. Past that point is where the roadmap turns into smoke; DRAM is expected to hit a wall somewhere in the 1x nm range, so 3D is the only option, but a variety of competing 3D technologies continue to cloud the industry’s direction. In addition to these scaling problems the server DRAM market is looking at transitioning from the socketed DIMM form factor to directly attached 3D chip stacks. Admittedly, the DRAM roadmap shows very comparable trends to the microprocessor manufacturing roadmap, with higher levels of package integration and 3D being the next step forward, but it still remains to be seen if DRAM can keep up with memory need of CPUs that are made on increasingly superior manufacturing processes.
This week Manuel Masiero and Achim Roos of Tom’s Hardware took a look at 15 microSDHC cards. They found that all of the cards they tested met or sometimes drastically exceeded the speed classifications that their manufactures gave them. With one class 4 card from SanDisk beating all of the higher rated class 10 cards in the sequential read metric, the testing shows that there’s significantly more to SDHC cards than the price tag and the class label. It’s nice to see manufacturers downplaying the performance of their products and delivering more than expected rather and the opposite for once.
Guru3D’s Hilbert Hagedoorn got a chance to review EVGA’s new GTX 560 Ti 2Win. With two GTX 560 Ti’s on a custom PCB and a 365 Watt power draw the 2Win is about 20 to 25 percent faster than Nvidia’s top single GPU card, the GTX 580. Its triple fan cooling solution and black fan shroud certainly give this card its own look. Pricing is an interesting story as the 2Win is $60 more expensive than a pair of GTX 560 Ti’s but only $20 more than a comparable GTX 580 from EVGA. As the successor to the GTX 460 2Win this will probably ignite old memories of other odd ball dual GPU cards like Sapphire’s 4850 X2, in any case it’s nice to see that the AIB vendors are still trying to spice up the wait for the 28nm GPU generation.
Nathan Kirsch of Legit Reviews played with AMD’s lowest end FX, the 4100, this week. He found the overclocking potential and ease of tweaking to be the largest reasons to purchase this chip. But he also found that the 4100 was bested by its socket FM1 cousins in a couple of benchmarks and offered less than notable stock performance. All things considered the reasons to buy the 4100 come down to its low price and its very good overclocking results. If this low-end FX review has shown anything it’s that AMD needs to drop power consumption, up per core performance, and do whatever it can to coax more overclocking headroom of its FX line. But the chances of any of that happening soon are pretty minimal. Oh B3 stepping, where art thou?S|A
Latest posts by Thomas Ryan (see all)
- AMD Launches its Radeon Instinct Server Accelerators - Jun 22, 2017
- A Look at AMD’s EPYC Line-up - Jun 21, 2017
- AMD Details EPYC’s Power Efficiency - Jun 20, 2017
- Intel’s Skylake-X and Kabylake-X Meet Our Scale - Jun 19, 2017
- MIPS for Self-Driving Cars with the I6500-F CPU - Jun 19, 2017