June 22, 20205 yr Commercial Member On 6/20/2020 at 6:31 AM, Prpn said: it is funny that LM has announced that they have now fixed it in their code after working together with AMD, How's that so funny? You didn't understand what I was saying in my posts and you're looking for some kind of disparity that way. Never mind though, because if the particular issue is resolved then that's good. Steve Waite: Engineer at codelegend.com
June 22, 20205 yr On 6/20/2020 at 7:31 AM, Prpn said: Hmm, for a problem that supposedly was entirely a fault on the side of AMD, it is funny that LM has announced that they have now fixed it in their code after working together with AMD, and no driver updates will be necessary. Seems to indicate no problems with the actual drivers or AMD would be implementing a fix on their side, just poor QC and testing on LM side as I indicated before. Could also be that LM, with help from AMD, has worked out an workaround so they don't have to wait for a driver fix. System: i9 [email protected] - 32 GB RAM - Aorus 1080ti --- Sim/Addons: P3D v5 + ProSim737
June 22, 20205 yr Commercial Member Without writing posts to cover all bases there's often someone that will look for disparity if they can, whether they understand the techniques or not: Looking back to the printer example which is perhaps easier to understand: Code does not target a printer (i.e. GPU) instead it looks at the capabilities of the printer and says in code something like this: 'if the border can be small than print to canvas BigLength x BigWidth, else if border must be large then print to canvas SmallLength x SmallWidth". So if the code does not check that border and canvas resolution properly then it might try to render to a canvas that cannot contain the image. Conversely, if the printer driver does not convey its capabilities correctly then the same problems arise. Therefor some problem with ascertaining GPU capabilities, the difference between AMD and Nvidia, could result in a crash. Whether that be the fault of LM or the GPU driver the fault exists and one or other is in need of fixing it. Sorry if that is too hard to understand for some. Just to cover all bases - Having said all that there can be many other reasons for a crash. It still does not mean that the code is targeting a GPU manufacturer. Steve Waite: Engineer at codelegend.com
Archived
This topic is now archived and is closed to further replies.