Which is what I was saying. DDR5 and the mobo pricing was already too high then about a year ago in a few weeks. Add another 50-120 for a decent air cooler or more for an aio and it would have been a very expensive build even on the low end. DDR5 price cuts didn't begin to hit their stride until around late march this year.Kinda wish they hadn't done that as well, but eh what can you do. If they had changed things up and made older coolers incompatible, they might have had even lower rates of early adoption.
One option is to reduce the IHS z height on Zen 5 and include an additional thin copper block with each CPU to fill that void without breaking compatibility with existing coolers.unless they can break compatibility which is doubtful.
Don't need a chiller to cool 7950X or a 7950X3D below their respective thermal targets, just a good block and enough radiator space. Both of mine have never hit the thermal target, even with PBO.Bit of a correction here. You're thermally limited but not truly. AMD said they designed the processor to use as much cooling as it can. In other words if you set this up under a loop and got the temps down through chilled water the processor will continuously boost until it hits a thermal wall again. The more cooling the more performance until the processor taps out what can be boosted. I personally think it's a weird design choice.
Insert my long old rant on the z axis height being a mistake because amd wanted to support a batch of coolers and save their consumers money.
My 280mm AIO has no issues cooling the 7950X. Liquid temps stay in the 20-30C range. The chip rarely reaches 95C, and usually when it does that is because the fans were too slow to ramp. Those are with 1500 rpm fans. If I wanted to get serious (and wanted to simulate riding in a jet), I have a pair of 3000 rpm fans I can install.Don't need a chiller to cool 7950X or a 7950X3D below their respective thermal targets, just a good block and enough radiator space. Both of mine have never hit the thermal target, even with PBO.
For air coolers and AIOs, and maybe crappy custom blocks or loops that run coolant temp too high, yeah. It's a thing.
It's going to be the same ( because it's still AM5 spec). Maybe AMD will change the socket in Zen6, if not, Zen7 for sure. I suspect they won't repeat their mistake.One option is to reduce the IHS z height on Zen 5 and include an additional thin copper block with each CPU to fill that void without breaking compatibility with existing coolers.
OK, a problem with my first B650 board. Running BOINC, and 2 8 thread tasks (no smt) its only taking 16% cpu per task and it should be 25%. But its running 5.3 ghz . WTF ???
What is the difference between Core Usage and Core Utility?I've found this most reliable tool to look into any windows load % related problem.
Its linux.....And it can't be accounting, its actually running way worse than all my 670e motherboards. Also, this is the first time its been stressed this much. This runs avx-512 and its for 10 days, all days (the challange)Can You post this part from HWinfo64 during that workload? I've found this most reliable tool to look into any windows load % related problem.
View attachment 85765
I never claimed you had to. I brought it up as an example of how Zen 4 will continuously try and boost and hit its thermal wall. Please try and read more carefully next time before jumping the gun.Don't need a chiller to cool 7950X or a 7950X3D below their respective thermal targets, just a good block and enough radiator space. Both of mine have never hit the thermal target, even with PBO.
For air coolers and AIOs, and maybe crappy custom blocks or loops that run coolant temp too high, yeah. It's a thing.
I'm not confident replying to this because I've never pasted a setup like this and am unsure how thermal transfer would work if you pasted the ihs and placed the copper block, pasted again with the cooler. You do have a floating piece of copper unless it was a moulded piece that went over the ihs but this is over complicated and opens the consumer to risks.One option is to reduce the IHS z height on Zen 5 and include an additional thin copper block with each CPU to fill that void without breaking compatibility with existing coolers.
Silly me to mention copper when there's the good ole graphite pad. Cheaper too. But it would have to be a thick pad. Not sure if thickness impedes graphite's ability to transfer heat. You then don't need any paste either.You do have a floating piece of copper unless it was a moulded piece that went over the ihs but this is over complicated and opens the consumer to risks.
I'm not sure either. I can see a third party company offering a fix not AMD. I do understand the oversight that went into amd's decision simply due to how expensive things were. We know Zen 4 was "late" to launch due to the pandemic, but I suspect had the pandemic not been a major issue it would have been even more expensive to launch earlier for end consumers. An extra 60-150 for an air cooler or aio would have made a nightmarish position for amd and as @DrMrLordX pointed out it would have lessened or outright killed any early sales amd got in the 3 weeks leadup to Raptor Lake releasing.Silly me to mention copper when there's the good ole graphite pad. Cheaper too. But it would have to be a thick pad. Not sure if thickness impedes graphite's ability to transfer heat. You then don't need any paste either.
you implied it simply had so much headroom that you can apply a custom loop with a chiller and it will keep boosting to hit the thermal target. That is clearly and blatantly false. The frequency and voltage limits built into the AGESA prevent it from simply scaling forever. It stops scaling very much far before a chilled water loop. Bad example. Don't blame me for bringing some reality to your hyperbole.I never claimed you had to. I brought it up as an example of how Zen 4 will continuously try and boost and hit its thermal wall. Please try and read more carefully next time before jumping the gun.
It was an example, and I was corrected. Your reply to me was thus superfluous. Why bother reposting what was already posted? Your not engaging in this discussion in good faith.you implied it simply had so much headroom that you can apply a custom loop with a chiller and it will keep boosting to hit the thermal target. That is clearly and blatantly false. The frequency and voltage limits built into the AGESA prevent it from simply scaling forever. It stops scaling very much far before a chilled water loop. Bad example. Don't blame me for bringing some reality to your hyperbole.
Another user has already stated they were able to allow it to boost to maximum without reaching thermal target with a good 360mm AIO.
You were not corrected. It was clarified the power and current limits still exist. I was pointing out that even with removing those limits, with a reasonably effective cooling solution it doesn't just simply "keep boosting". You're wrong. Deal with it.It was an example, and I was corrected. Your reply to me was thus superfluous. Why bother reposting what was already posted? Your not engaging in this discussion in good faith.
Clarification is correction in this case. If you clarify a misnomer it is a correction. The correction of my wrong statement. Again, please read carefully.You were not corrected. It was clarified the power and current limits still exist. I was pointing out that even with removing those limits, with a reasonably effective cooling solution it doesn't just simply "keep boosting". You're wrong. Deal with it.
A chilled custom water loop is way, WAY more than what I have going myself, or the other user's 360mm AIO.
It wasn't a misnomer. It was an outright falsehood. Deal with the fact you were wrong.Clarification is correction in this case. If you clarify a misnomer it is a correction.
Reported and blocked.Your aggressive behaviour is way out of line. You're tripping over your words and lack of comprehension.It wasn't a misnomer. It was an outright falsehood. Deal with the fact you were wrong.
AM5 will just be Zen4 and Zen5?It's going to be the same ( because it's still AM5 spec). Maybe AMD will change the socket in Zen6, if not, Zen7 for sure. I suspect they won't repeat their mistake.
Zen 5 is allegedly coming out the first half of 2024, and I expect the 3d variants to come out in 2025. AMD never said they wouldn't release lower end or non X processors a year after those. I don't really like how AMD is vague when it comes to support. Reminder, this is the company that relented zen 3 capability on older motherboard after extreme community backlash.This article says AMD confirms AM5 will extend into 2026+:
Was 2025 last I knew. Article says:AM5 will just be Zen4 and Zen5?
This article says AMD confirms AM5 will extend into 2026+:
AMD Confirms AM5 Socket Will Extend to 2026, Zen 5 to Use RNDA 3.5 GPUs
A new roadmap slide from AMD contains several intriguing updates to its future CPU and GPU plans.www.extremetech.com
When are you expecting Zen6 and Zen7 to be released?
Though several nuggets of info are contained in a new slide straight from AMD, the most prominent is that the company has extended AM5's lifespan by one year, pushing it out to 2026.