A couple of days ago, NVIDIA released a BETA BIOS (which you can download here) for its nForce 680i SLI motherboards that are currently being sold by EVGA and BFGTech. Normally we don't write about BIOS updates, but this one is quite substantial.
If you're a regular reader of our Hardware & Overclocking Forum, you may have come across a few threads about EVGA's nForce 680i SLI motherboard, where we've outlined our rather poor experiences with three different EVGA boards from three different batches.
The third board was specifically selected and pre-tested by NVIDIA, but still didn't fix the problems that we have had with the board(s).
Many end users had experienced SATA issues and data corruption. Now, we are not quite sure whether that's what we've experienced, but we are treating our situation (although with slightly different issues than what most are reporting) as a SATA-related issue.
We are now in the process of wiping any of the hard drives that have been used on one of the EVGA nForce 680i SLI motherboards we've attempted to test and starting from completely fresh (even our drive containing the main backup of our benchmarking suite is being wiped clean). We're doing this so that there is no question that any of the data on our drives is corrupt. Thankfully, we have a backup of our benchmarking suite on our office network, and also on DVD.
After over three weeks of exhaustive testing, NVIDIA was no closer to solving our own problems, since many of the issues we were having were different to what other reviewers and end users had experienced and passed back to NVIDIA and EVGA. However, we had started to come to the conclusion that many of the hard locks and crashes we were experiencing were actually related to the board being set up too aggressively. In particular, we felt the chipset timings (that aren't exposed in the BIOS) were the source of the problems because we've had similar issues in the past that turned out to be related to the board being set up too aggressively.
Interestingly, today, HardOCP has had the chance to sit down with Drew Henry, General Manager of MCP Business at NVIDIA, in order to discuss the issues that many people have had, along with detailing how the BIOS engineers worked to identify and fix the issues that were apparent.
It turns out that our diagnosis was on the right lines, as Drew referred to the issue as a problem with the signal timings on the board. Since nForce 680i SLI uses the same MCP as the AMD version of nForce 590 SLI - and to the best of our knowledge, there are no data corruption issues with that MCP - the issue is actually related to the way that the memory controller handled data (i.e. very badly).
If you're a regular reader of our Hardware & Overclocking Forum, you may have come across a few threads about EVGA's nForce 680i SLI motherboard, where we've outlined our rather poor experiences with three different EVGA boards from three different batches.
The third board was specifically selected and pre-tested by NVIDIA, but still didn't fix the problems that we have had with the board(s).
Many end users had experienced SATA issues and data corruption. Now, we are not quite sure whether that's what we've experienced, but we are treating our situation (although with slightly different issues than what most are reporting) as a SATA-related issue.
We are now in the process of wiping any of the hard drives that have been used on one of the EVGA nForce 680i SLI motherboards we've attempted to test and starting from completely fresh (even our drive containing the main backup of our benchmarking suite is being wiped clean). We're doing this so that there is no question that any of the data on our drives is corrupt. Thankfully, we have a backup of our benchmarking suite on our office network, and also on DVD.
After over three weeks of exhaustive testing, NVIDIA was no closer to solving our own problems, since many of the issues we were having were different to what other reviewers and end users had experienced and passed back to NVIDIA and EVGA. However, we had started to come to the conclusion that many of the hard locks and crashes we were experiencing were actually related to the board being set up too aggressively. In particular, we felt the chipset timings (that aren't exposed in the BIOS) were the source of the problems because we've had similar issues in the past that turned out to be related to the board being set up too aggressively.
Interestingly, today, HardOCP has had the chance to sit down with Drew Henry, General Manager of MCP Business at NVIDIA, in order to discuss the issues that many people have had, along with detailing how the BIOS engineers worked to identify and fix the issues that were apparent.
It turns out that our diagnosis was on the right lines, as Drew referred to the issue as a problem with the signal timings on the board. Since nForce 680i SLI uses the same MCP as the AMD version of nForce 590 SLI - and to the best of our knowledge, there are no data corruption issues with that MCP - the issue is actually related to the way that the memory controller handled data (i.e. very badly).
MSI MPG Velox 100R Chassis Review
October 14 2021 | 15:04
Want to comment? Please log in.