*Official Benchmark Thread* - Post it here or it didn't happen :D

[Mr. Fox]

2 minutes ago, tps3443 said:

Yeah I’m running chilled water for sure. But it’s not that cold. I stick to 15C daily, and 15C in the runs below. But the chip isn’t delidded so it still gets very hot. If it was direct die I’d be whooping on my old 13900KS which I can’t lol. That chip could do 6.1P/4.8 simply because it had the thermal headroom being direct die and 15C water. I already touch 80C temps at only 350 watts. So, my cooling isn’t that good. The block and contact is actually pretty poor I think. Simply because It’s a stock chip with stock Intel soldered IHS so temps aren’t good like DD+Chiller. 
 

These are old runs on my last OS.
 

uploading pictures

As best I can tell from your screenshots, your VID and VCore load values are nearly identical to mine. About 1.270V under full tilt load.

[tps3443]

12 minutes ago, Mr. Fox said:

As best I can tell from your screenshots, your VID and VCore load values are nearly identical to mine. About 1.270V under full tilt load.

Didnt you just post a run the other day at 6Ghz pulling sub 400 watts? That 550 seems inaccurate. Maybe I’m crazy lol. That seems like a lot of power though. 
 

 

[Mr. Fox]

24 minutes ago, tps3443 said:

Didnt you just post a run the other day at 6Ghz pulling sub 400 watts? That 550 seems inaccurate.

Correct. It may not be accurate, but it is what the BIOS is telling HWiNFO64. I don't have any control over that. I do remember posting that, but it might have been when the CPU was still in the mini-ITX motherboard. I can't recall now.

 

For perspective, here is the white Apex with the chiller cooling the water to 16°C. Note the wattage. This is the SP112 CPU so it pulls 1.300V under load versus the SP117 only needing 1.270V.

 

Other than the extra 0.030V needed, the rest of the BIOS settings are the same as the Encore without chilled water. Power draw is the same (according to HWiNFO64). Can't tell you if that is accurate or not, only what it reads. And, Benchmate gets the info from HWiNFO64.

[electrosoft]

3 hours ago, tps3443 said:

Nahhh I’m too lazy for all that. My 8400c36 profile is stable right now, and I don’t wanna deal with re-stabilizing it after clearing bios. 

 

I've been there before lol...

 

21 minutes ago, tps3443 said:

No, I hit like 350 watts for 6.0P/4.8E/5.2R manual voltage R23. It hits about 45,200. The 600 watts is really astounding. 

 

15 minutes ago, Mr. Fox said:

Not sure how you are able to do that other than with chilled water. As mentioned, under load my CPU is drawing 1.270V.

 

14 minutes ago, tps3443 said:

Yeah I’m running chilled water for sure. But it’s not that cold. I stick to 15C daily, and 15C in the runs below. But the chip isn’t delidded so it still gets very hot. If it was direct die I’d be whooping on my old 13900KS which I can’t lol. That chip could do 6.1P/4.8 simply because it had the thermal headroom being direct die, lapped die, and 15C water. I already touch 80C temps at only 350 watts. So, my cooling isn’t that good. The block and contact is actually pretty poor I think. Simply because It’s a stock chip with stock Intel soldered IHS so temps aren’t good like DD+Chiller.
 

 

I think that's one reason I've asked to go get a standard 360mm/480mm AIO to come back to reality in the past and watch some of your chips come crashing back down to earth. 🙂 The only way you will truly know how your chip stacks up under completely normal conditions is to run it as such and not outlier setups with chilled water / custom loops and delided/DD.

 

You're a lot closer this time around @tps3443 having not delidded your chip (yet) but you're still running on custom water chilled. Plunk down $100-150 and keep a 360mm/480mm AIO around for a reality check. An added bonus is you can also compile "real world / normal" results to pit against your chilled awesomeness. You would have a whole set of "OOB on a standard AIO" before vs "Chilled / DD" after (and all the progression steps in between). You would show not only good samples of silicon but how they shine in every scenario. Heck, pick up a couple of air coolers too to really flesh out the results!

 

In summary, there's just no way you can downplay the importance of custom / chilled water in any capacity when presenting benchmarking runs. It really is a game changer.

 

17 minutes ago, Papusan said:

Regarding the massive power consumption I think different AC/DC loadline values play some role into this. I mean the power consumption may being somewhat scewed. 

 

 

For the pict lover @electrosoft😀. A new baby on the way to Papusan's plaza. The N570GTX Twin Frozr III Power Edition/OC. Or GTX 570. This will be my replacement card for my dead GTX570 (my hobby you know🤪). This MSI nvidia card offer Vcore (+150mV), Memory (+50mV) and Aux voltage (+30mV). This is not what you see nowadays. Is old better than  new bro @ryan😁

 

Look at the beauty here.... https://www.finn.no/bap/forsale/ad.html?finnkode=289263887&ci=8

 

 

MSI Introduces GeForce GTX 570 TwinFrozr III Power Edition Graphics Card

PRESS RELEASE by Apr 14th, 2011

 

I wouldn't mind seeing a return to that look again! I love the grey metal and that black back ground for the MSI logo with the red. They can update and use this design anytime!

[Mr. Fox]

35 minutes ago, Papusan said:

Regarding the massive power consumption I think different AC/DC loadline values play some role into this. I mean the power consumption may being somewhat scewed. 

Yes, that is definitely likely as well. Following the thread at Overclock.net, if I tune my LLC the way most of them are suggesting (which I did on the Encore) so that the VID and VCore match under load my power draw goes way up. It may not be accurate, but tuning it that way may cause the sensors to have false readings.

 

If I put AC and DC LLC at 0.01 each the power draw readings are like 100W less and the VID is higher than the VCore under load.

[tps3443]

So the power usage is wrong. That makes sense now. I was wondering like what in the world man. You’re gonna weld the Direct die to the die. 😂

[Mr. Fox]

1 hour ago, tps3443 said:

So the power usage is wrong. That makes sense now. I was wondering like what in the world man. You’re gonna weld the Direct die to the die. 😂

I think TurboV Core just killed the SP112 13900KS in my white Apex. 🤢

 

I launched it and tried to move the voltage a bit and when I clicked apply the system powered off and now I get Q-code 00 and the red CPU light comes on. I did the BIOS flashback from a USB drive after clearing CMOS did nothing. It flashed successfully and when I powered back on it still has 00 and red light.

 

This really sucks. This is the second 13900KS that has died when installed in an ASUS motherboard. In fact, every CPU that I have owned that died, if memory serves me correctly, was installed in an ASUS motherboard when it died. My 10900KF golden sample died while installed in a Apex.

 

Edit: I guess I am not the only one. I think in this case it killed the guy's CPU and mobo. 

 

https://rog-forum.asus.com/t5/intel-700-600-series/asus-turbo-vcore-killed-my-z690-asus-apex-and-12900k/td-p/874202

 

I hate ASUS.

 

I guess I will try the CPU in another system and see if it boots.

[Papusan]

1 hour ago, Mr. Fox said:

I think TurboV Core just killed the SP112 13900KS in my white Apex. 🤢

 

I launched it and tried to move the voltage a bit and when I clicked apply the system powered off and now I get Q-code 00 and the red CPU light comes on. I did the BIOS flashback from a USB drive after clearing CMOS did nothing. It flashed successfully and when I powered back on it still has 00 and red light.

 

Hmmm☹️ Very sorry bro Fox.  Maybe there is something into this Asus Turbo Vcore killed my Z690 Asus Apex and 12900k

 

Edit. Didn't see you have already posted the link. 

 

I only use it for changing BCLK. And Asus TurboVcore is buggy as well.

 

Edit. GTX 970

https://hwbot.org/submission/5481410_papusan_3dmark11___performance_geforce_gtx_970_21495_marks?recalculate=true

[electrosoft]

26 minutes ago, Mr. Fox said:

I think TurboV Core just killed the SP112 13900KS in my white Apex. 🤢

 

I launched it and tried to move the voltage a bit and when I clicked apply the system powered off and now I get Q-code 00 and the red CPU light comes on. I did the BIOS flashback from a USB drive after clearing CMOS did nothing. It flashed successfully and when I powered back on it still has 00 and red light.

 

This really sucks. This is the second 13900KS that has died when installed in an ASUS motherboard. In fact, every CPU that I have owned that died, if memory serves me correctly, was installed in an ASUS motherboard when it died. My 10900KF golden sample died while installed in a Apex.

 

Edit: I guess I am not the only one. I think in this case it killed the guy's CPU and mobo. 

 

https://rog-forum.asus.com/t5/intel-700-600-series/asus-turbo-vcore-killed-my-z690-asus-apex-and-12900k/td-p/874202

 

I hate ASUS.

 

I guess I will try the CPU in another system and see if it boots.

 

Hopefully it boots in another system. In that scenario, I think I'd prefer the motherboard to be faulty not the CPU as the motherboard still has the MFG warranty on it.

 

At least it wasn't the SP117 golden chip.

 

Good luck!

 

 

[Mr. Fox]

1 hour ago, Papusan said:

 

Hmmm☹️ Very sorry bro Fox.  Maybe there is something into this Asus Turbo Vcore killed my Z690 Asus Apex and 12900k

 

Edit. Didn't see you have already posted the link. 

 

I only use it for changing BCLK. And Asus TurboVcore is buggy as well.

 

Edit. GTX 970

https://hwbot.org/submission/5481396_papusan_3dmark11___performance_geforce_gtx_970_21393_marks?recalculate=true

 

55 minutes ago, electrosoft said:

 

Hopefully it boots in another system. In that scenario, I think I'd prefer the motherboard to be faulty not the CPU as the motherboard still has the MFG warranty on it.

 

At least it wasn't the SP117 golden chip.

 

Good luck!

 

 

I dropped my Celeron test CPU into the Apex and it is running, so it is the SP112 CPU that it killed. The SP108 is in the mini-ITX system. I wish it would have been that one. The SP112 has a 120 P-core SP rating the same as the SP117.

 

ASUS sucks. I wish it would have been the motherboard as well. Easier to replace than a nice CPU sample.

 

I will try putting the 13900KS back into it again tomorrow and pray for a miracle. If no luck I guess I will have to waste money on a 14900K that would not have been necessary and hope it's not a worse silicon sample.

[Etern4l]

7 hours ago, Papusan said:

Todays joke? 

 

Nvidia... Not much love for 20xx series RTX cards. The paying guinea pigs for first gen RTX pushed down the drain for now. Nvidia didn't even bother mention 2000 series RTX. Hmmm, just lovely.

 

NVIDIA unveils ‘Chat with RTX’: AI Chatbot powered locally by GeForce RTX 30/40 GPUs
https://videocardz.com/newz/nvidia-unveils-chat-with-rtx-ai-chatbot-powered-locally-by-geforce-rtx-30-40-gpus

In this instance there is a reason. In order to run on small GPUs, the models have to be dumbed down and only Ampere onwards support the required low res modes. Still a good marketing gimmick I guess, given the limitations.

[Mr. Fox]

1 hour ago, Papusan said:

I only use it for changing BCLK. And Asus TurboVcore is buggy as well.

I bet you won't use it any more. 😞 You can bet your booty I will never use or install it ever again. And, you know the goons at ASUS would never in a million years accept responsibility for it. When the Z490 Apex died and killed the 10900KF they replaced the motherboard and blew me off on the CPU it killed. Wasn't their problem. Intel paid for the CPU (refunded my purchase price because none were in stock to replace it). I wonder if that is what killed the last one in the Z690 Apex. I had TurboV Core on that system as well.

[Papusan]

2 hours ago, Mr. Fox said:

I bet you won't use it any more. 😞 You can bet your booty I will never use or install it ever again. And, you know the goons at ASUS would never in a million years accept responsibility for it. When the Z490 Apex died and killed the 10900KF they replaced the motherboard and blew me off on the CPU it killed. Wasn't their problem. Intel paid for the CPU (refunded my purchase price because none were in stock to replace it).

 

I guess I will try get a BCLK button hardware tool. Better than use this half baked Asus software mess. Will you ask for replacement for the dead Cpu? At least the MB survived. That's better than nothing I guess.

 

4 hours ago, Mr. Fox said:

This really sucks. This is the second 13900KS that has died when installed in an ASUS motherboard. In fact, every CPU that I have owned that died, if memory serves me correctly, was installed in an ASUS motherboard when it died. My 10900KF golden sample died while installed in a Apex.

 

The sad part... There ain't much options for OC motherboards. Most sucks. I hope we see more 2 dimm board for next gen. Maybe MSI will come up with an equal as the Unfy-X and not this new 2 dimm baby OC boards for baby PC boxes.

 

GTX970 (forgot add screenshoot of 3DM11 Extreme for Hwbot. That would be 4th place). Oh'well, I get a new chance 🙂

https://hwbot.org/submission/5481425_papusan_3dmark11___entry_geforce_gtx_970_33907_marks?recalculate=true

 

https://hwbot.org/submission/5481429_papusan_3dmark___cloud_gate_geforce_gtx_970_69697_marks?recalculate=true

 

https://hwbot.org/submission/5481444_papusan_unigine_superposition___1080p_xtreme_geforce_gtx_970_2727_points?recalculate=true

 

https://hwbot.org/submission/5481441_papusan_3dmark___time_spy_extreme_geforce_gtx_970_2365_marks?recalculate=true

 

https://hwbot.org/submission/5481434_papusan_3dmark___time_spy_geforce_gtx_970_5146_marks?recalculate=true

 

[Rage Set]

3 hours ago, Mr. Fox said:

I bet you won't use it any more. 😞 You can bet your booty I will never use or install it ever again. And, you know the goons at ASUS would never in a million years accept responsibility for it. When the Z490 Apex died and killed the 10900KF they replaced the motherboard and blew me off on the CPU it killed. Wasn't their problem. Intel paid for the CPU (refunded my purchase price because none were in stock to replace it). I wonder if that is what killed the last one in the Z690 Apex. I had TurboV Core on that system as well.

 

I am sorry bro. Whenever the infamous Code 00 appeared for me, it meant the mobo died. Count it as another learning experience when you decide in the future to pick up an Asus board. 

 

I am at the point that I would pick up a Gigabyte mobo over Asus if I were forced to choose, and that's saying something.

[Mr. Fox]

8 hours ago, Rage Set said:

 

I am sorry bro. Whenever the infamous Code 00 appeared for me, it meant the mobo died. Count it as another learning experience when you decide in the future to pick up an Asus board. 

 

I am at the point that I would pick up a Gigabyte mobo over Asus if I were forced to choose, and that's saying something.

I would if it were a two-DIMM mobo, but I won't purchase another four-DIMM motherboard unless it is HEDT and each stick is a discreet memory channel. The two per channel sucks. For some reason there are hardly any two slot enthusiast mobos, which is idiotic because those with four suck at memory overclocking and are the exact opposite of an enthusiast product. The Tachyon is like the Unify-X. They made enough for people to want one, then quietly vanished. Seems almost like a publicity stunt.

 

The code 00 has generally meant a dead mobo for me as well. The last one was the Z490 Apex and when the mobo died it took that 10900KF with it. The 13900K that died in the Z690 Apex has code 55, so I am guessing the mobo did something to fry the memory controller on the CPU. Code 55 is memory not detected. Just for giggles, I tried that 13900K and it still shows code 55 in whatever mobo it is installed in.

 

I'd even entertain trying another AssRock mobo again if it were a two-DIMM model versus going with ASUS again the next time. ASUS is just a sucky company that makes overpriced half-assed garbage. 

 

I ordered a 14900K on Amazon. I hope it's a decent silicon sample. It will be here tomorrow. I really didn't want to waste $500+ on a relabeled 13900KS.

 

@tps3443what's an average or better P/E/MC SP rating for 14900K? What are your SP ratings @Papusan and @Talon?

[Etern4l]

How much do people expect to pay for a GPU water block? I was checking out the options for my 3090 Ti FE and all I could find so far was a £300+ unit from Extra Kostly Water Blocks (out of stock anyway).

[electrosoft]

10 hours ago, Mr. Fox said:

 

I dropped my Celeron test CPU into the Apex and it is running, so it is the SP112 CPU that it killed. The SP108 is in the mini-ITX system. I wish it would have been that one. The SP112 has a 120 P-core SP rating the same as the SP117.

 

ASUS sucks. I wish it would have been the motherboard as well. Easier to replace than a nice CPU sample.

 

I will try putting the 13900KS back into it again tomorrow and pray for a miracle. If no luck I guess I will have to waste money on a 14900K that would not have been necessary and hope it's not a worse silicon sample.

 

Ouch, hope this doesn't turn out like your rash of bad luck a year or two ago when you had multiple parts die and we all put distance between ourselves and you so the bad juju didn't transfer. 🙂

 

4 hours ago, Mr. Fox said:

I would if it were a two-DIMM mobo, but I won't purchase another four-DIMM motherboard unless it is HEDT and each stick is a discreet memory channel. The two per channel sucks. For some reason there are hardly any two slot enthusiast mobos, which is idiotic because those with four suck at memory overclocking and are the exact opposite of an enthusiast product. The Tachyon is like the Unify-X. They made enough for people to want one, then quietly vanished. Seems almost like a publicity stunt.

 

The code 00 has generally meant a dead mobo for me as well. The last one was the Z490 Apex and when the mobo died it took that 10900KF with it. The 13900K that died in the Z690 Apex has code 55, so I am guessing the mobo did something to fry the memory controller on the CPU. Code 55 is memory not detected. Just for giggle, I tried that 13900K and it still shows code 55 in whatever mobo it is installed in.

 

I'd even entertain trying another AssRock mobo again if it were a two-DIMM model versus going with ASUS again the next time. ASUS is just a sucky company that makes overpriced half-assed garbage. 

 

I ordered a 14900K on Amazon. I hope it's a decent silicon sample. It will be here tomorrow. I really didn't want to waste $500+ on a relabeled 13900KS.

 

@tps3443what's an average or better P/E/MC SP rating for 14900K? What are your SP ratings @Papusan and @Talon?

 

With a 14900KS possibly on the horizon, you could have waited but I don't know if I could sit there and stomach a dead system due to a CPU.

 

With the 14900k binning as it has been, pray for an SP100+ chip, expect a sub SP100 that will be worse than your SP117. Things get decent >=SP102+ for 14900k chips. MC ratings have routinely proven to be the most unreliable for 14th gen.

 

 

 

 

[Mr. Fox]

3 hours ago, Etern4l said:

How much do people expect to pay for a GPU water block? I was checking out the options for my 3090 Ti FE and all I could find so far was a £300+ unit from Extra Kostly Water Blocks (out of stock anyway).

I won't pay the EK GPU block premium or the Optimus premium. Not worth it at all. You're paying for a brand name and slightly different aesthetics and gaining little or nothing in terms of results unless you buy something that was poorly designed and below average to begin with. In that case you're making major sacrifices just to save a buck, which is also not a respectable option. If you're going to make sacrifices to save a few bucks, then just stick with the stock air cooler and spend nothing.

 

For me it needs to be about $200 or less or it represents poor value. The Bykski blocks and the Alphacool block I just purchased for the Gigabyte 4090 have all been under $180.

44 minutes ago, electrosoft said:

Ouch, hope this doesn't turn out like your rash of bad luck a year or two ago when you had multiple parts die and we all put distance between ourselves and you so the bad juju didn't transfer. 🙂

 

With a 14900KS possibly on the horizon, you could have waited but I don't know if I could sit there and stomach a dead system due to a CPU.

 

With the 14900k binning as it has been, pray for an SP100+ chip, expect a sub SP100 that will be worse than your SP117. Things get decent >=SP102+ for 14900k chips. MC ratings have routinely proven to be the most unreliable for 14th gen.

I hope it does not turn out that way. That would really suck. 

 

I've really viewed the 14900K as a meaningless sideshow if you own a decent 13900KS. That being said, I am not willing to pay an extra $200 for a brand new 13900KS at this point. The 14900K makes good sense as an upgrade from something less than a 13900KS, but not really an intelligent purchase for good bin 13900KS owners. If the 14900KS costs as much or more than the 13900KS did, I will skip the priviledge being offered to throw money in the toilet and just wait until I can buy a used one that has a known good SP rating for less money. Better to buy a nice used car and let someone else take the loss on depreciation. I apply the same logic to CPUs now.

 

I forgot that I had a $200 credit waiting for me at NewEgg when I ordered the 14900K from Amazon last night, so I ordered one from NewEgg after requesting to cancel the Amazon order. Now Amazon said it is too late, so it seems two will be arriving tomorrow. I guess I will check which one is better and return the one that is inferior. Unless I somehow get lucky on both, in which case I might just keep both and sell the SP108 13900K or save it as a spare. The delidded SP117 would run nice and cool in the mini-ITX setup. IceMan Cooling also has LGA 1700 bare die frames now, so I might even be able to run it bare die with the AIO.

 

Funny observation. The Celeron max core temp under a stress test on the Apex with the big custom loop was 28°C. The min/max core temps were within 2°C of idle temperature, LOL.

 

[Papusan]

6 hours ago, Mr. Fox said:

@tps3443what's an average or better P/E/MC SP rating for 14900K? What are your SP ratings @Papusan and @Talon?

 

Probably 60mv worse than your KS. Or equal 1-2 bin lower possible OC headroom for me. But benching from 6.0GHz up to 6.5GHz depending on benchmarks in 2D and 3D. It's SP94 with 102P-cores. It does it's job but with too much voltage. And its' not an MC lottery winner either as @electrosoft nicely point out. MC-SP74. Tragic. But again... My benchmarks is still in the upper end vs those compete with golden samples. I'ts not about the tool but how you use it🙂 Most of my benchmark scores is quite nice and are up in the upper end.

 

2 hours ago, electrosoft said:

With the 14900k binning as it has been, pray for an SP100+ chip, expect a sub SP100 that will be worse than your SP117. Things get decent >=SP102+ for 14900k chips. MC ratings have routinely proven to be the most unreliable for 14th gen.

 

2 hours ago, Mr. Fox said:

For me it needs to be about $200 or less or it represents poor value. The Bykski blocks and the Alphacool block I just purchased for the Gigabyte 4090 have all been under $180.

 

EK... 3 times more. They are disgusting https://www.ekwb.com/shop/ek-quantum-vector-fe-rtx-3090-ti-d-rgb-black-special-edition

 

Alphacool https://shop.alphacool.com/en/shop/gpu-water-cooling/nvidia/13056-alphacool-eisblock-aurora-acryl-gpx-n-rtx-3090ti-founders-edition-with-backplate

 

Alphacool Eisblock Aurora Acryl GPX-N RTX 3090TI Founders Edition with Backplate below $90 USD vs EK and +$290 @Etern4l

 

 

[tps3443]

@Mr. Fox I’d test both 14900K’s and settle for whichever can produce about 1.300-1.320V under load stock at 6Ghz. Because once you pop the IHS and run direct die it’ll take off another 0.040-0.050 more under load putting you at or below 1.270V equivalent to your KS. I do not think the Bios VFcurves are comparable between 13900KS and 14900K. So I wouldn’t go by that too much. 
 

I do think SP97+ is pretty dang good though. Many of these lower SP chips are on direct die 6Ghz+ doing 1.243V-1.260V load. So I think you will be surprised. 
 

Also, check the E-Cores and ring as well. Both on 14th gen are pretty amazing. Unless it’s a complete dud chip lol.

[electrosoft]

23 minutes ago, tps3443 said:

@Mr. Fox I’d test both 14900K’s and settle for whichever can produce about 1.300-1.320V under load stock at 6Ghz. Because once you pop the IHS and run direct die it’ll take off another 0.040-0.050 more under load putting you at or below 1.270V equivalent to your KS. I do not think the Bios VFcurves are comparable between 13900KS and 14900K. So I wouldn’t go by that too much. 
 

I do think SP97+ is pretty dang good though. Many of these lower SP chips are on direct die 6Ghz+ doing 1.243V-1.260V load. So I think you will be surprised. 
 

Also, check the E-Cores and ring as well. Both on 14th gen are pretty amazing. Unless it’s a complete dud chip lol.

 

I wouldn't keep an SP98 if it's my hard earned money buying it. I'd want an SP100+ and prefer an SP102+ but as you say DD/Chiller can close the gap on a lot of variance but using that logic I guess even an SP93 is a keeper since you're going to pop the top and chill it down lol.

 

 

 

 

[tps3443]

1 minute ago, electrosoft said:

 

I wouldn't keep an SP98 if it's my hard earned money buying it. I'd want an SP100+ and prefer an SP102+ but as you say DD/Chiller can close the gap on a lot of variance but using that logic I guess even an SP93 is a keeper since you're going to pop the top and chill it down lol.

 

 

 

 

I’m actually thinking about down sizing my whole system. I’m moving soon, and I’m not sure if I want to drag all this with me lol. I’m considering getting out of the “extreme game” and just running normal case with a AIO on a normal CPU with stock speeds. It’s been fun. But definitely I’m not looking forward to moving my work PC 🤣

 

 

[Etern4l]

9 minutes ago, tps3443 said:

I’m considering getting out of the “extreme game” and just running normal case with a AIO on a normal CPU with stock speeds.

 

Definitely need to get some betting going around that resolution lol

AIOs suck for 13900K+ at "stock speeds", you would be back on a custom loop inside a month.

[Mr. Fox]

3 hours ago, electrosoft said:

MC ratings have routinely proven to be the most unreliable for 14th gen.

Honestly, the MC SP rating is not all that meaningful on 13th Gen. Among the 13900K/KS chips I have had they ranged from SP78 to SP87 and all of them overclocked the memory stable using the same timings and voltage, so unless you are trying to run 9000+ with tight timings I don't know that the MC SP rating is that critical. I can't see any behavioral difference in the MC SP78 than I do with the SP87. Since I had not intended to buy a 14900K at this point in time I have not paid a lot of attention to know if the MC SP rating matters more on 14th Gen than it does on 13th Gen.

 

As often as I see variance in SP ratings for the P- and E-core randomly on both of my ASUS motherboards I don't have a huge amount of trust in their importance. I've seen my SP ratings sudden tank for no reason while the CPU is running the same, then powering down and switch to a different BIOS chip it goes back to normal. I may have to flash up or down to a different firmware version and then back again for the SP ratings to return to their previous values, but in the meanwhile nothing changed with the behavior or performance of the CPU in conjunction with an erroneously reported SP rating. 

12 minutes ago, Etern4l said:

 

Definitely need to get some betting going around that resolution lol

AIOs suck for 13900K+ at "stock speeds", you would be back on a custom loop inside a month.

There is definitely some truth to that. With an AIO you get an inferior waterblock cold plate, comparatively very weak pump, very low flow rate and minuscule coolant volume. Better than air cooling by a little bit, but no comparison to a custom loop with a well engineered cold plate, strong pump, high flow rate and much, much larger volume of coolant. If they made an AIO that worked the same as a custom loop it would take up a whole lot more space and be so expensive that you would be better off building a custom loop instead.

 

[tps3443]

4 minutes ago, Mr. Fox said:

Honestly, the MC SP rating is not all that meaningful on 13th Gen. Among the 13900K/KS chips I have had they ranged from SP78 to SP87 and all of them overclocked the memory stable using the same timings and voltage, so unless you are trying to run 9000+ with tight timings I don't know that the MC SP rating is that critical. I can't see any behavioral difference in the MC SP78 than I do with the SP87. Since I had no intended to buy a 14900K at this point in time I have not paid a lot of attention to know if the MC SP rating matters more on 14th Gen than it does on 13th Gen.

 

As often as I see variance in SP ratings for the P- and E-core randomly on both of my ASUS motherboards I don't have a huge amount of trust in their importance. I've seen my SP ratings sudden tank for no reason while the CPU is running the same, then powering down and switch to a different BIOS chip it goes back to normal. I may have to flash up or down to a different firmware version and then back again for the SP ratings to return to their previous values, but in the meanwhile nothing changed with the behavior or performance of the CPU in conjunction with an erroneously reported SP rating.

MCSP does not impact how far a memory controller will overclock, it directly impacts how much IMC voltage is needed to achieve the same overclock. 
 

You may find a 14900K mcsp75 and it runs DDR5 8200 Memtest stable with 1.475V CPU IMC. But you could test a 14900K mcsp85 that only needs 1.300v CPU IMC voltage for the same 8200. 

I think higher mcsp directly leads to lower cpu IMC voltage, and also lower SA voltage too.