MX150 Variants Not All Equal
There was a lot of talk about nVidia’s Max Q GPUs for Notebooks and Ultrabooks. Keeping cool and quiet by optimizing various things, well it turns out that the optimizations were nothing more than lying about the specs of the GPU and hiding it behind one name. In this case there are two variants of MX150 GPUs, one with higher Core Clocks and Memory Clocks and one with Lower Clocks. The problem is they are not distinguishing which one is which and calling them all MX150’s but that can lead to differences of around 20-25% in their performance.
This is a problem of branding but one they do not seem keen to point out readily. They should have different names and indeed they do have the Max Q title for this purpose but it is not being used and the only way to tell is to check the actual specs with something like GPUZ or a good review site before buying a laptop with some variant of the MX150. This is a simple mistake, I am sure but it comes off as incredibly shady considering they are two completely different GPUs in their capabilities. It would be like Intel not distinguishing betweek the normal CPUs and the K series chips or AMD selling their CPUs with out pointing out the X or non X versions and selling them all as the same name.
Since there does not seem to be any identical laptops with both variants there is not telling right now if they are trying to pull one and charge more the the MX150 compared to the MX150 (yes deliberately confusing on purpose). They should be named MX150 Standard and MX150 Max Q, or MX150 and Again MX150 Max Q, or even MX150 and MX140 or some other naming convention to differentiate them.
The other thing is, and this is certainly deliberate, is that to a regular consumer Max Q would sound like the better version. Max is good right it means More! When in reality if they did make the effort or put the proper name on then the unknowing consumer would be getting a worse experience. Marketing… Woo.
Titan V Miscalculations
This one is a little less clear if this is real problem or not but nVidia do confirm that there is at least one application that has inconsistent results when using the Titan V, Amber (a molecular dynamics package). I am way over my head on this one so i will resort to quotes from the article.
One engineer told The Register that when he tried to run identical simulations of an interaction between a protein and enzyme on Nvidia’s Titan V cards, the results varied. After repeated tests on four of the top-of-the-line GPUs, he found two gave numerical errors about 10 per cent of the time. These tests should produce the same output values each time again and again. On previous generations of Nvidia hardware, that generally was the case. On the Titan V, not so, we’re told.
An industry veteran, who alerted us to the issue, reckoned this is due to a memory issue. Chip companies normally push their high-end silicon to the limit to maximize performance. Nvidia may be overclocking or red-lining its Titan V in some way, causing read errors from memory. These mistakes are carried forward in calculations, resulting in numerical errors. Another cause could be a design blunder.
There was no issue with its Titan X card based on its Pascal architecture, we’re told.
Updated to add
A spokesperson for Nvidia has been in touch to say people should drop the chip designer a note if they have any problems. The biz acknowledged it is aware of at least one scientific application – a molecular dynamics package called Amber – that reportedly is affected by the Titan V weirdness.
“All of our GPUs add correctly,” the rep told us. “Our Tesla line, which has ECC [error-correcting code memory], is designed for these types of large scale, high performance simulations. Anyone who does experience issues should contact [email protected].”
This would seem a fairly silly problem to have with this particular card considering it is aimed at calculation workloads and high accuracy tasks suck as simulations and not Gaming like many of the previous Titan cards.
Video for those interested, it has other topics too besides the ones above.