But nobody uses cl42. As I said, cl30 out of the box for EXPO is VERY common.26 cycles is well out of spec for 6000, JEDEC is 42.
JEDEC is for boring office PCs or workstations.26 cycles is well out of spec for 6000, JEDEC is 42.
Whatever the heck happened to their RAM timings? Are boring slower speed kits now available at CL26 or are they using the expensive kits and downclocking those to 5200/5600 MT/s?AMD Ryzen 9 9900X3D im Test
Der Ryzen 9 9900X3D hat nur 6 Kerne mit 3D V-Cache. Wie schnell ist er im Vergleich zu 9800X3D und 9950X3D mit 8 und als „9600X3D“ im Test?www.computerbase.de
Unless specified otherwise they use stock settings for both frequency and latency, when the 9000 serie was released there were already fast RAMs well above 5600MT/s but they used CL32 timing, so they updated their benchs surely due to mature BIOSes but i know that they have very fast RAMs since they did ocked tests up to 7200-8000MT/s.Whatever the heck happened to their RAM timings? Are boring slower speed kits now available at CL26 or are they using the expensive kits and downclocking those to 5200/5600 MT/s?
at 1080P. Do people game at that resolution anymore ?
So I guess you guys would love this then. TLDR: 9800X3D sucks and is barely faster than 5800X3D.Especially with a 5090, and using DLSS?
Theyre trying to exacerbate the load on memory to get a performance difference. Seems like this goes right into the bin with the rest of reviews that test in unrealistic configurations for the sake of "isolating variables".
Edit: they also tested with a 9950X, which also exacerbates the memory load. What 5090 owner isnt rocking a 9800X3D or 9950X3D?
Ah yes, classic. Take our criticism of one extreme, hyperbolic scenario by suggesting an opposite, equally extreme, hyperbolic scenario.So I guess you guys would love this then. TLDR: 9800X3D sucks and is barely faster than 5800X3D.
I mean, come on...
Really, when testing CPU performance you want to test CPU performance. If it turns into GPU benchmark, what's the point? I don't know what you think is realistic scenario, but I don't know what kind of relevant data you going to get out of it.Ah yes, classic. Take our criticism of one extreme, hyperbolic scenario by suggesting an opposite, equally extreme, hyperbolic scenario.
I don't know what your point is supposed to be. I made clear what mine was - They went out of their way to create the most artificially memory-constrained gaming configuration they could to get exacerbated results and I doubt the real-world applicability of those results for several reasons I already stated.Really, when testing CPU performance you want to test CPU performance. If it turns into GPU benchmark, what's the point? I don't know what you think is realistic scenario, but I don't know what kind of relevant data you going to get out of it.
My tech budget is going to be out of control this year. 🤣So they've tested single rank 2x16GB kits. Great? How does this compare to dual rank 2x32 or 2x48GB 6000C30 and 6000C28 kits (and the coming soon 2x48 6000C26 kits)
I have older games that cap out at 1080p.at 1080P. Do people game at that resolution anymore ?
I don't know what your point is supposed to be. I made clear what mine was - They went out of their way to create the most artificially memory-constrained gaming configuration they could to get exacerbated results and I doubt the real-world applicability of those results for several reasons I already stated.
Your logic here is impeccable.It's like trying to test soil quality effects on plant growth but not giving the plants any water. Congratulations on proving that soil quality doesn't matter I guess.
And it turned out that WE ended up under water with this thread.I was worried that the thread was going to end up under water
9900X3D review with simulation of a 9600X3D by disabling a CCD :
AMD Ryzen 9 9900X3D im Test
Der Ryzen 9 9900X3D hat nur 6 Kerne mit 3D V-Cache. Wie schnell ist er im Vergleich zu 9800X3D und 9950X3D mit 8 und als „9600X3D“ im Test?www.computerbase.de
Its a 12 core. Why do you say 6 core ????? Its 24 threads/cores.AMD should not make 9900x3d as it's really just 6-core cpu - thus inferior to all other x3d models. Or if they want to make 12-core version at least they should do it asymmetric maintaining fully functional 8-core x3d die. Or at least sell that 6-core dual die chip at lower price than fully functional die parts.
Its a 12 core. Why do you say 6 core ????? Its 24 threads/cores.
and where did you read that ? about the thread director ?For it's purposed use case -gaming AMD thread director disables 3d-cacheless chiplet. So it's only 6-core and inferior cpu to not only 9950x3d but also 9800x3d which have full 8-core x3d die. 9900x3d should not exists as it's lower grade cpu with higher price than 9800x3d. Only people that doesn't know about their relative performance will buy one - so AMD should stop being ass and just stop selling such a fraudulent product.
and where did you read that ? about the thread director ?
There is no fraud. Independent reviews of this product exist.AMD should stop being ass and just stop selling such a fraudulent product.
It is not inferior if your workloads do not make use of more than 6 cores. (Including workloads which nominally use more than 6 cores but don't actually saturate them, for example because they are predominantly bottlenecked by performance of a single central program thread.)it's really just 6-core cpu - thus inferior to all other x3d models.
How about a link please. Or it doesn't exist.AMD does not hide it - x3d cpu's should park 3d-cacheless CCD when running games - and performance will suffer if it doesn't. So AMD has instructions how to monitor core parking at gaming.
For those reasoning 5600x is not inferior to 5800x. But it is if workload scales beyond 6 cores - so AMD values those cpus right - 5600x is cheaper than 5800x. 9900x3d ain't valued right - it only have 6 core x3d die and pretty useless other 6 core die - for gaming that is inferior to 9800x3d and should be priced so. Users only buy that cpu over 9800x3d because they though it's better as it have more cores - if they look any review they can see that it ain't.It is not inferior if your workloads do not make use of more than 6 cores. (Including workloads which nominally use more than 6 cores but don't actually saturate them, for example because they are predominantly bottlenecked by performance of a single central program thread.)
How about a link please. Or it doesn't exist.