You’re not reading it wrong, that’s really what they’re asserting. GN is saying that it is an active problem resulting in the failures.
GN is sometimes really good and sometimes really bad. They do tons of great data-driven benchmarking and research like this, and then jump to bizarre conclusions at the end, or make bizarre proclamations like “long term value doesn’t exist because that’s not how money works”.
For a while back in 2019 they were declaring the death of 6-core CPUs because of 0.1% lows, and their supporting evidence basically was a couple obviously-broken games like BFV and Far Cry 5 where a 2C4T pentium was outscoring their 6C processor by a factor of 2 in the 0.1% lows… and they didn’t see a problem even when it was raised etc.
Another example was Steve insisting last december that “nvidia was leaving the gaming industry” because of “recent” quotes from Jensen about nvidia being an AI company now… and the citation literally on screen behind him dated the quotation to the mid-2010s. Raised it to him, doesn’t care. He got his clicks.
Like, as the gaming community prepares for the launch of the 5090 sometime this fall or early next year… how’s that working out?
It’s not uncommon for steve to do this. Like there’s certainly an issue here, and oxidation is an issue that existed in the past. Steve thinks therefore that it must be the issue for all these chips, even though intel says that’s basically a small problem that was fixed over a year ago. Steve disagrees, on the basis of evidence such as:
(This space left intentionally blank)
Like people need to get it through their heads that reviewers are humans too, with their own foibles and weaknesses. Some articles he does incredible investigative work. Other ones he’s arguing that we need to abandon 6C processors in favor of (checks notes) stock 2C pentiums because of frametimes, or that spending more can never deliver long term value if it lasts better because “that’s not how money works” (exasperated Steve face). It all depends whether he’s being data driven or whether he’s engineering a support for an argument on a given day - he sucks way more when he’s working backwards to fit data to conclusion.
(ps: my other favorite thing about that chart is the frametimes getting worse as processor frequency increases. The 8600k and 9600k are literally the exact same silicon, same cache, same core count. So a pair of processors both lost frametime consistency as they were overclocked… and actually the faster overclock lost more frametime consistency than the slower overclock. It’s literally a straight correlation between “higher clocks = worse frametimes” even on the exact same chip models, which obviously indicates an engine problem, along with the previous point about pentiums being faster than a processor with 3x the cores and 50% more threads at double the frequency on the same architecture. Like it’s just a facially absurd result to stand as the primary peg in the “6c are dead now!” argument. And the rest of the games weren’t much better - BFV was not in a healthy state in 2018-2019, for example, it’s not really indicative of anything except the state of frostbite engine in 2018.)
GN is sometimes really good and sometimes really bad. They do tons of great data-driven benchmarking and research like this, and then jump to bizarre conclusions at the end, or make bizarre proclamations like “long term value doesn’t exist because that’s not how money works”.
For a while back in 2019 they were declaring the death of 6-core CPUs because of 0.1% lows, and their supporting evidence basically was a couple obviously-broken games like BFV and Far Cry 5 where a 2C4T pentium was outscoring their 6C processor by a factor of 2 in the 0.1% lows… and they didn’t see a problem even when it was raised etc.
https://i.imgur.com/Py69p36.jpg
https://gamersnexus.net/images/media/2018/cpus/2600k/intel-i...
Another example was Steve insisting last december that “nvidia was leaving the gaming industry” because of “recent” quotes from Jensen about nvidia being an AI company now… and the citation literally on screen behind him dated the quotation to the mid-2010s. Raised it to him, doesn’t care. He got his clicks.
https://youtu.be/VSSb-t76EpU?t=147
Like, as the gaming community prepares for the launch of the 5090 sometime this fall or early next year… how’s that working out?
It’s not uncommon for steve to do this. Like there’s certainly an issue here, and oxidation is an issue that existed in the past. Steve thinks therefore that it must be the issue for all these chips, even though intel says that’s basically a small problem that was fixed over a year ago. Steve disagrees, on the basis of evidence such as:
(This space left intentionally blank)
Like people need to get it through their heads that reviewers are humans too, with their own foibles and weaknesses. Some articles he does incredible investigative work. Other ones he’s arguing that we need to abandon 6C processors in favor of (checks notes) stock 2C pentiums because of frametimes, or that spending more can never deliver long term value if it lasts better because “that’s not how money works” (exasperated Steve face). It all depends whether he’s being data driven or whether he’s engineering a support for an argument on a given day - he sucks way more when he’s working backwards to fit data to conclusion.
https://www.youtube.com/watch?v=tu7pxJXBBn8&t=273s