6200 AGP version FSB1200 for nForce5 60,000 suspect AMD CPUs seized

Janvier 2005 L M M J V S D 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 You can read previous news using our archives. [ Print all news of this day ]

<< Previous day's news December 30, 2004

Next day of news >> January 4, 2005

6200 AGP version   Posted on 03/01/2005 at 13:41 by Marc

	Palit is the first one to release GeForce 6200 AGP on the market. It is the non TurboCache version, the NV43-V which is originally a castrated NV43 (6600). Once more the HSi PCI Express to AGP NVIDIA bridge is used to include this PCI Express graphic chip on an AGP graphic card. Like the standard 6600, the 6200 is clocked at 300 MHz and 275 MHz for the DDR 128 bits memory. However the number of pixel pipeline is reduced from 8 to 4. performances are then strongly reduced and instead of reaching 50 fps with a 6600 with Far Cry in 1024*768 the 6200 only reach 36 fps (the 6600 reach the same framerate… but in 1280*1024). In fact the GeForce 6200 performances are comparable to a 9600 Pro. If the price gap between the 6200 and 6600 AGP is equivalent to the price gap between the 6200 and 6600 PCI Express (approximately $20), we recommend you to choose the 6600.

FSB1200 for nForce5   Posted on 03/01/2005 at 13:14 by Marc

	Several fellow colleagues seems to be amazed by the fact that the future Chipset NVIDIA for Pentium 4 (nForce5 ?) could work via Overclocking to a bus speed of 300 MHz, equivalent to a « FSB1200 », as the Pentium 4 bus is quad pumped (4 words transmitted per cycle). But it isn’t extraordinary at all as the nForce5 is supposed to support officially the FSB1066 required for the last Pentium 4 Extreme Edition. So a FSB1200 only correspond to a 12.6% Overclocking. The only chipset Intel which support the FSB1066, the i925XE, also works with such a frequency. In our article on the i925XE we were even able to reach 320 MHz in synchronous mode and 335 MHz with memory in asynchronous mode. The nForce 5 should be release during the first quarter 2005.

60,000 suspect AMD CPUs seized   Posted on 03/01/2005 at 13:05 by Marc - source: DigiTimes

	Based on tips provided by AMD, the Taiwanese police Friday raided an electronics company located in Taiwan, and seized a total of 60,000 suspect AMD CPUs. The suspect AMD CPUs, were defective CPUs that would normally have been destroyed, but that might have been stolen from one of AMD’s plants for re-marking and sold on the grey market. According to the Liberty Times magazine, one million of remarked AMD processors at least would have been shipped in Asia and Europe. This information hasn’t been confirmed by AMD. Processors sold on the grey market aren’t necessarily faulty processors. We have published one new on remarked processors last November 2003 (in French). Back then, AMD Athlon XP 2400+/2600+ processors were actually remarked Duron 1.4 GHz. This modification was possible because Duron only had 64 KB of cache out of 256 KB activated. By changing manually AMD K7 processors’ bridges, it was possible to change the size of the cache activated but also the multiplying coefficient, the FSB or even the electrical tension. Today it is possible to find fake Athlon 2800+ with 512 Kb of cache. These processors are based on Athlon XP « Thorton » (cf. this news in French), which has 512 KB of cache but only 256 KB activated (the last 256 aren’t necessarily functional). As you may have noticed, forgers made efforts to improve characters used for the processor reference. However, L2 bridge modifications (for the cache) remain visible. The first bridge which was cut has been refilled to activate the 512 Kb of cache. The only way to avoid this type of surprise is to choose AMD Socket A box versions rather than OEM’s. If you buy one processor the L2 bridge verification will help you to see whether if it is a remarked version. Because of the absence of bridges and because of the metallic plate, it isn’t possible to remark AMD Socket 754/939 (Athlon 64).