Question Zen 4 builders thread

[Markfw]

I know its early, but as someone pointed out, I better get my ram now before the scalpers are all over it.

To go with a 7950x, what do you think of this ram:

https://www.newegg.com/g-skill-32gb-288-pin-ddr5-sdram/p/N82E16820374369?Item=N82E16820374369

(edited for my best recommendation, 12/5/2023)

 

[Chaotic42]

Anyone in here feeling bold?

 

[Markfw]

Anyone in here feeling bold?

No, and I have no need to. -25 co optimizer.. fixes it all

 

[Chaotic42]

I enabled PBO and got my linpack up to 764.6396. Still below 40C. I bet someone who knew what they were doing could get this thing to scream.

 

[Timmah!]

Between 420 AIO cooling CPU and 240 GPU AIO, which one would you mount upfront and up top? The case would be full-tower, Fractal Design 7XL.

Can i mount the 3 supplied 14cm fans (with case) upfront and then mount the 240 GPU rad onto them, creating sort of push-pull config? Or that would be only possible if the supplied fans were 12cm too?

 

[Det0x]

Some binning stats for the 7950x

LLC 5, L3 idle temp ~20 degrees, max load ~60 temp (~185w PPT with 6000/2000 memory)
1200mv vcore set, ~1110mv get under load

Maximum stable 3x in a row Cinebench r23 allcore clocks:
7950x SP 115 (SP 120 + SP 110) = 5400/5300mhz
7950x SP 117 (SP 121 + SP 112) = 5500/5200mhz
7950x SP 116 (SP 119 + SP 114) = 5450/5350mhz
7950x SP 116 (SP 120 + SP 111) = 5425/5325mhz

Trying to find a good sample before going direct die

 

[DrMrLordX]

Trying to find a good sample before going direct die

Depending on your workload, you'll either want the best possible CCD0 or the best average SP between CCD0 and CCD1.

For gaming I'd pick sample #2. For applications or for a high-res streamer box I'd pick sample #3.

 

[igor_kavinski]

7950x SP 115 (SP 120 + SP 110) = 5400/5300mhz
7950x SP 117 (SP 121 + SP 112) = 5500/5200mhz
7950x SP 116 (SP 119 + SP 114) = 5450/5350mhz
7950x SP 116 (SP 120 + SP 111) = 5425/5325mhz

So AMD is doing this on purpose? Pairing one good CCD with another not-that-good CCD? Why not release a slightly more expensive 7990X with both high quality binned CCDs?

 

[DrMrLordX]

So AMD is doing this on purpose? Pairing one good CCD with another not-that-good CCD? Why not release a slightly more expensive 7990X with both high quality binned CCDs?

They've been doing it since Matisse. The logic is simple: use CCD0 in lightly-threaded workloads when the maximum boost clock actually matters. Use CCD1 only in MT-heavy scenarios when the boost algo dials back clocks to stay within PPT limits. It allows AMD to sell more high-end parts since they don't need two "good" CCDs per package.

 

[igor_kavinski]

It allows AMD to sell more high-end parts since they don't need two "good" CCDs per package.

I can't say that I'm OK with that. At least with Intel's KS, you know that you are getting the very best of their silicon.

 

[DrMrLordX]

I can't say that I'm OK with that. At least with Intel's KS, you know that you are getting the very best of their silicon.

Functionally, there would be no difference unless you put a 7950X (or 7900X) under extreme cooling and tried locking both CCDs at high clocks with static vcore.

 

[In2Photos]

I can't say that I'm OK with that. At least with Intel's KS, you know that you are getting the very best of their silicon.

Is this not the same as P core and E core though? 2 sets of cores running at different frequencies?

 

[Markfw]

Is this not the same as P core and E core though? 2 sets of cores running at different frequencies?

Is this not a Zen 4 BUILDERS thread ???? No Raptor crap allowed here.

 

[In2Photos]

Is this not a Zen 4 BUILDERS thread ???? No Raptor crap allowed here.

Chill out Mark, good grief. I was simply stating that AMD is doing nothing wrong with having 2 CCD with different capabilities, just using an analogy. Lighten up!

OK, not trying to rain on your parade, as that sounds much more efficient than stock. But my 7950x's are running 142 watt, and the benchmarks are almost the same as at stock.

My point is, that both Raptor Lake and Zen 4 can run with far less wattage than stock, but Zen 4 appears to have the lead in efficiency still.

It's not like you have ever posted in an Intel BUILDERS thread about AMD right?

 

[igor_kavinski]

2 sets of cores running at different frequencies?

I was just wondering why AMD won't introduce their own KS equivalent with both good CCDs. Maybe TSMC has more trouble with consistent die quality than Intel.

 

[In2Photos]

I was just wondering why AMD won't introduce their own KS equivalent with both good CCDs. Maybe TSMC has more trouble with consistent die quality than Intel.

Wouldn't the KS typically be a gamer type chip? So X3D would be it's direct competition?

 

[DrMrLordX]

I was just wondering why AMD won't introduce their own KS equivalent with both good CCDs. Maybe TSMC has more trouble with consistent die quality than Intel.

Because, by default, it wouldn't be worth it. They would have to raise the PPT to get both CCDs to clock high enough in MT workloads for anyone to notice a difference.

 

[moinmoin]

Because, by default, it wouldn't be worth it. They would have to raise the PPT to get both CCDs to clock high enough in MT workloads for anyone to notice a difference.

At least with AM5 the TDP headroom is there so AMD could do some Intel KS style binning now.

 

[igor_kavinski]

Wouldn't the KS typically be a gamer type chip? So X3D would be it's direct competition?

KS has higher all core boost so it benefits apps too.

 

[DrMrLordX]

At least with AM5 the TDP headroom is there so AMD could do some Intel KS style binning now.

They could but

eww

 

[Markfw]

OK, got my new cheap heatsink for my 7950x build (finally a real build post....) Its a 250 tdp, so it will handle my co optimized 142 watt just fine, and it was $50 shipped ! I don't have it fired up, but it was way easier than my dark rock pro 4 or a nhd15

 

[eek2121]

Some binning stats for the 7950x

LLC 5, L3 idle temp ~20 degrees, max load ~60 temp (~185w PPT with 6000/2000 memory)
1200mv vcore set, ~1110mv get under load

Maximum stable 3x in a row Cinebench r23 allcore clocks:
7950x SP 115 (SP 120 + SP 110) = 5400/5300mhz
7950x SP 117 (SP 121 + SP 112) = 5500/5200mhz
7950x SP 116 (SP 119 + SP 114) = 5450/5350mhz
7950x SP 116 (SP 120 + SP 111) = 5425/5325mhz

Trying to find a good sample before going direct die

Please do not rely on ASUS SP-ratings as an indicator of silicon quality. That rating is not an indicator of silicon quality and may be removed or changed in a future release.

EDIT: As an example, ASUS had incorrect secondary voltages for various areas in early BIOS revisions. When they corrected the issue, the SP ratings for each core on my chip changed completely.

On top of that, at least one member of the press is seeing very odd things happen on ASUS boards vs. AsRock. They are awaiting boards from other vendors to pinpoint the issue.

So AMD is doing this on purpose? Pairing one good CCD with another not-that-good CCD? Why not release a slightly more expensive 7990X with both high quality binned CCDs?

They have done this since Zen 2.

 

[alexruiz]

OK, got my new cheap heatsink for my 7950x build (finally a real build post....) Its a 250 tdp, so it will handle my co optimized 142 watt just fine, and it was $50 shipped ! I don't have it fired up, but it was way easier than my dark rock pro 4 or a nhd15
View attachment 71745

What cooler is it?
DeepCool AK620?

 

[Markfw]

What cooler is it?
DeepCool AK620?

Yes, after I fire it up, I will reply as to how its working/

 

[Timmah!]

I can't say that I'm OK with that. At least with Intel's KS, you know that you are getting the very best of their silicon.

it definitely hurts my OCD, not gonna lie

 

[biostud]

Yes, after I fire it up, I will reply as to how its working/

Supposedly it should be better than both dark rock and noctua.