Problem in encryption AES-256 for SandForce

SandForce SF - 2000 controllers are affected by a problem in the management of the hardware encryption in AES 256 bit.

An internal audit conducted by LSI discovered a concern in the management of the hardware encryption offered by the SandForce SF-2000 series chips. Other solutions of hardware encryption for hard disks, SandForce controllers use two systems separate encryption, first of 256 bit and 128 bit a second. According to the audit, the first level 256 bit is not present, not working in 128-bit mode.

For individuals, the problem is not really embarrassing, but for companies and Governments that have special constraints on the encryption of hard disks, it becomes a little more serious. For example, the US Government asked that used hard drives are encrypted in AES 256 bit, what these controllers offered natively. However, if the controller does not handle...

If this is really disturbing, some manufacturers of DSS that uses its default controllers offer rebate offers or alternatives. For example, Intel offers fully reimburse customers who have purchased flash drives of the series 520 - Yes, no error on the press release - 1 July, and have just contacted customer service before October 1 for it. Kingston for its SSDNow V200 + and its KC100 and adds that a solution is being developed jointly with LSI to repair this problem of encryption. Of course, other manufacturers using the SandForce controllers will also publish their news releases to meet the demands of their customers.

Ivy Bridge and temperature: a problem of TIM confirmed

TIM (for Thermal Interface Material) is a thin layer of thermal paste between the nucleus and the IHS. Intel would have used a somewhat limited TIM which results in a rise in temperature important in overclocking on the Ivy Bridge.

Our Japanese colleagues from PC Watch had fun to disassemble a third-generation Intel Core, Core i7-3770K CPU to be exact, and analyzed the impact of the thermal paste between the nucleus and the IHS (Integrated Heat Spreader) used by Intel. As we have seen in testing of overclocking the processor by our expert Thomas "Pt1t" Kerrinckx, the heater is much more important than expected. The emitted doubts writing is confirmed by tests of Japanese and used thermal paste is not the most effective.

In the design of a processor, the caster has two solutions, it pastes the IHS on the ring with a thermal paste either it directly welded the IHS on the ring. In both cases, the heat emitted by the nucleus must be evacuated at best to the heatpipe itself intercalated with a thin layer of thermal paste. The image of PC Watch above illustrates this point, with black kernel, in red the two layers of pulp thermal (the bottom being posed by Intel and the top by the user) and clear grey the IHS on the dark grey that represents the base of installed heatpipe.

Our Intel Core i7-3770K decapsulation...

Thus, the Overclocking can engage in the gurgling on this generation of processors to then put a new layer brand, as the paste OCZ Freeze Extreme or the Coollaboratory Liquid Pro. The results are without appeal and the performance of the pulp used by Intel inhibits the potential of OC of Ivy Bridge. The decapping operation is not without risk still and the gain may be minimal if the processor is not of good quality. However, PC Watch tests show a difference of 15 to 20 ° C reference processor and the modified processor, which allows up to 5 GHz to 1.55 V without any configuration plant during the Cinebench 11.5 multi-threaded test in the case of Liquid Pro.

                  

A table and a record of temperatures by PC Watch

Remains to be seen if revisions of these third-generation Intel processors will correct this problem with the pulp of better quality, or with an IHS directly welded to cut short just this controversy... But nothing is less sure.