I don't really see the point or recording "scores" for synthetic benmchmarks .... kinda like seeing how fast you could swim at a depth of 10,000 feet. Or how long I could hold my breathe on a planet with 100% oxygen atmosphere. Since this is not something you would ever do in the course of any day ever for work or entertainment, what's the point ? Furmark is an extremely useful tool for determining the capacity of your cooling system. We use it to limit pump speeds at the point where additional flow provides no reasonable gains. Same for cooler and rad fans ... utilities are simply tools and the most important step is choosing the right tools for the the task you want to accomplish. You want todial in CPU / GPU cooling systems, use P95 and Furmark ... you want to determine stability, use RoG Bench and Unigine / 3D mark Tools.
CPU - I use P95 (old version) to cycle my CPU temps up to mid 80s because w/o that, I can't break out of the mid 70s .... my OC is limited by voltage, BIOS says 1.3875 but this spikes up to 1.44 when Open CL is present and up over 1.5 for a millisecond now and then when AVX is present. Once I cycle the TIM 4 -5 times up to 85C and back down to room temps, the build will never see P95 again. Furmark also works well for the same task on GPUs ... after mounting a water block and applying TIM on 25 or more separate surfaces, it loses its spreadability at room temps and the quality of your work is much better where ya beganthe 1st one than at the end when ya finished the 25th one. By leaving the rad fans off or at fixed low speed, I can cycle those temps up to > 80 and back rather quickly. On this box, that was 4 years ago and they haven't seen anything > 42C since.
We have been building Hi end CAD PCs since 1991 (gaming boxes around 2006) and creating component lists for custom built laptops for about 18 years. A component doesn'fail because it saw the orange flame on the Furmark logo. It dies because the particuar card in question was permitted to exceed it's capabilities. Have used Furmark / OCCT or something like it on every single build we have ever done since it 1st came out. We sit at the monitor and watch voltage and temps ... prepared to stop the test if and when we see anything of concern. If you stay away from cards with known issues, that shouldn't be a concern. If you stop the test when you see a temp you don't like or the particular voltage limit for that card, than there's no concern. We have never had a fried card and such a result is completely outside our range of experience.
That does not mean it can not occur. People have died using camping stoves, people have died using chainsaws. I have been using camping stoves since boy scouts in the 1960s and chanisaws since I had a fireplace. I'm not dead yet. To avoid such an outcome ... 1) Stay away form products with known issues .... EVGA 570 SC is a good example, 970 SC when 1/3 of the heat sink missed the GPU and it wa smissing thermal pads ..... or 1060 - 1080 SC, also missing pads ... yea we all saw the pics. Or, if ya wanna use it... raise settings gradually and know when you are approaching limits. But then there's 2) .... Pay attention. You are using a tool where if the product has a flaw and you do not use it responsibly, you can damage something. But that goes for any tool. Im thinking of the guy who used his wet vac to vacuum up a gasoline spill in his garage. Do we ban wet vacs ? or do we just figure that it's Darwin's rule at work here ?
Lastly, when I contacted nVidia about a laptop that wouldn't run OCCT, they said "just use Furmark".
Yeah keep on using the furmark virus!
