Jump to content
NotebookTalk

electrosoft

Member
 View Profile  See their activity

  • Posts

    2,799

  • Joined

  • Last visited

  • Days Won

    129

 Content Type 

Everything posted by electrosoft

  1. electrosoft
    My systems in their standard setup when working on them (open and ready). Left my old faithful Corsair Carbide Air 540 that is used as my "open bench" test rig. Right now it has an Asus Z590-A Prime using to test and bin 10900k's for the X170SM-G and some 11900k's when they sporadically come my way. To the right is the Phanteks Enthroo Pro II housing an Asus Z690 Strix D4 w/ 12900ks (most likely going to go frugal and switchback to my 12900k), Strix 3080 and stretching 2x8GB G.Skill sticks to 4266 G1 atm fishing for errors.
  2. electrosoft
    Much better than m SP81 and SP82 I had and the SP82 was actually a decent CPU. Once you get in there and do some auto stock runs at 4.9 -> 5.x (wherever it craps out), it will let you know what you're working with pull and temps wise. Could be a decent chip with a good IMC.
  3. electrosoft
    Congrats on the purchase and I am VERY interested in your results and assessment especially CPU thermal characteristics and PL along with a touch of CCC, XTU or TS to see how much it can be tamed and get a look around that BIOS.
  4. electrosoft
    ~10% drop in temps (73->66) = 7c drop. Fans are running slightly less. Maybe that little bit of headroom will let you push a higher OC than before for some benchmarks. See where it separates itself from old numbers >5.2 all core.
  5. electrosoft
    Considering the X170 is running dual 280w PSUs I'm inclined to agree its PL won't be as robust as previous DTRs. Binning definitely can play a major role in laptop thermals. I can't imagine sticking a poorly binned 12900k that not only pulls 240w+ at stock but is leaky to boot especially when you're working with the thermal and power limitations of the X15 versus the robust expectations of a desktop. I'm running a Silicon Lottery 10900k in my X170SM-G at the moment and under the same conditions it is pulling ~20w less than even the original binned ZtecPC 10900k and about 13 degrees cooler plus the ZtecPC 10900k is delidded and I'm running the SL 10900k as is. Delidded with a good Rockitcool or BartX (preferred) IHS I could bring that down even more. Being able to park the E-cores is a step in the right direction. I know Dsanke was able to partially unlock the stock Insyder BIOS for the X170 series. While it isn't anywhere near full tilt Prema on the X170SM-G it definitely gave you more options and control. I'm just really curious because my 12900k (sitting in a box atm while I run a 12900ks through the mill) is an exceptionally cool running 12900k. When i tested it against the other 3 12900k's I tried and saw the temps and pull my first thought was, "Pity Clevo isn't offering a 12th gen laptop. This would be perfect." With the wide variance of temps and pull of the 12900k, which is on par with the 11900k variance, binning a chip for use in this would be a high priority. I checked RJ and nothing but I didn't message them either to see if they're going to carry it as a barebones. I'm not sure if XMG is going to carry it? It is in many ways. But in the right hands the BIOS can most likely at least be unlocked a bit. If Prema was so inclined he could get in there and really fine tune it since it uses the Insyder or we could get a Dsanke partial unlock. I'm not sure what they can do about the lack of mxm given the current state of mxm. I'm curious what model of 3070ti this is since that timespy GPU score is sub 2080 Super stock. It almost feels like a hybrid model of a true DTR and a T&L but outside of the soldered GPU everything else looks like classic Clevo.
  6. electrosoft
    Yikes. Well, let's see how it pans out under real testing both bins and the IMC. I had an SP82 that was performing above an SP89 with a better IMC and clocks. What are the P and E cores and V/F curve?
  7. electrosoft
    The only knock on the AC LF II is the auto pump control that scales automatically depending on load. I wish they had separated the pump for its own header to leave it running at max if so desired. Of course the upside is one header controls everything. You might be like me and at the point of starting to slowly shift mentally to a custom loop. I don't ever see myself going full Mora, DD, chiller, etc... but a custom loop does provide much higher flow rates and bigger and better blocks for even better cooling. As for the case, I use a Phanteks Enthroo Pro 2 and the Corsair 7000D has 13lbs extra on it (yikes). Size wise they're about the same. I'm a gym rat too but I can imagine that 7000D loaded up and moving it around won't be fun. I'm a big proponent of if you're not happy return it. I'm not going to limp along with something that isn't providing the paid value for my uses and send it back from whence it came.
  8. electrosoft
    Awesome! Good luck!
  9. electrosoft
    This. Tweezers and jeweler's glasses did it for me along with a pair of the wife's cosmetic small scissors.
  10. electrosoft
    I hope 90w PL1 isn't the max as that is basically a mobile level PL. I can't see that being the max or even close for a desktop chip which even spec'd from Intel is 125w which is rarely adhered to. All previous Clevo Intel DTRs of the last 6 generations (6th -> 11th) had much much higher PLs ranging from 150w up to 200w. If you're looking to bin you will need a 12900 / 12900k(f) that doesn't so much overclock well put happens to have fantastic stock pull and thermals and can take a good UV. I've binned 4 12900k chips and the stock pull running CB23 as a baseline has ranged from 241w down to 208w right out of the box. Applying undervolts and/or reducing LLC had the 208w chip drawing down to 150w up to 162w (depending on heat) under load. Even my new SP94 12900ks can't get close to that bottoming out at ~173w set to 4.9/3.9 like a 12900k. I usually qualify chips on a desktop board (Asus normally) that lets me differentiate between which chips work best at stock and can run full tilt under simulated conditions. It's the same approach I took to hunt down good candidate 11900k and 10900k to work in the X170KM and X170SM so they wouldn't throttle and handle full runs. Let's say the BIOS is locked down. You put a better chip in there it will draw less by default. If XTU (or TS) works, you can then lower than even more with a UV. I would love to know the BIOS options. Coming from Eurocom (and hopefully who else offers it), I'd like to see some CPU control. Disabling e-cores would be great. We don't know how much XTU comes into play (or CCC) to apply an undervolt too. It is using the Insyde BIOS the same as the X170 series so it must likely can be unlocked to a degree. Looking at those benchmarks, especially Timespy, that system is scoring lower than my X170SM-G both in GPU and CPU multi and of course looking at the CB23 score that single thread score is righteous and multi is better. It looks to ship with a single 230w PSU so that is troubling as it touts supporting up to the 12900k. We definitely need more real world testing and info to see what it can and can not do and how its power distribution works in CPU only, GPU only and combination scenarios.
  11. electrosoft
    Last run before switching back to my SP91 12900k for some comparative tests. 12900ks SP94 can definitely do 5.4 all core but it will need a delid and better cooling to really make it shine as I'm thermal throttling on my 420mm AIO. Wish it was winter right now. A delid at the minimum will get it doing 5.4 all core on the 420mm AIO easy. Better cooling will just kick it up a notch to maybe 5.5. Core #1 is clearly the monster of the group.
  12. electrosoft
    Run versus pass real sustained stress tests are two different things. The rigor and qualifications Intel expects for 5.2 all core is much more rigorous than a lot of our "soft stable" settings we accept. That article was very much on point especially here: "Is the Core i9-12900KS ‘just’ a machine-binned 12900K? Yes, and in my opinion, it doesn't need to be anything else. Let me show you what a good chunk of them can do on standard water cooling and follow that up with multiple world records and gold trophies with liquid nitrogen cooling, no big deal." Tighter tolerances, expectations and performance. That is what you're paying for. Even a poorly binned 12900KS will still run 5.2 all core....hot and heavy sure but it will do it; just like the 12900k will follow the same track except 300mhz slower. I've just been running mine with AI OC enabled while I put it through its single core and WoW paces and seeing it handle 5.7ghz in wow on 2-3 cores no problem even raiding for hours on end which hits the CPU the hardest in WoW, temps sitting in the 60s after hours of play tickles my fancy. My SP91 12900k couldn't do it. Not even close. Watching it hit 2200+ single core in CB23 really shows you the architectural gap between it and 11th and especially 10th gen. Still, I can be frugal at heart so I am going to swap back in the 12900k SP91 later this week to see if the 12900ks is really worth the $831.66 it cost. My return window is till May 30th so it isn't like I don't have time. 🙂 One thing this 12900ks can't do is touch my SP91 at 4.9 settings. I can dial it down to ~174w CB23 well above the sub 160w the SP91 draws but I'll need to retest it with 1403. Agreed. this is what a delidded and relidded 10900k looks like with a BartX after I did it a few days ago with the nickel coated BartX. I do love the look of his IHS's. Temps are fantastic. For reference 12900ks SP94 letting Asus control the entire CPU narrative (for better or worse) AI Optimized, MCE Enabled, No AI BIOS control w/ +2 TVB enabled:
  13. electrosoft
    I really like the refinements RC made to the 12th gen model in regards to the holding cap and pressure plastic screw. The core unit looks about the same though. They need to do a nickel plated model like BartX to avoid dealing with the LM soak/hardening issues. You also hear Alex (his side kick in this one) even talk about XTU and how this one is better binned than the last one they tested. Tale of two 12900ks chips: I just got back earlier from my bud's place to unfurl his new Best Buy 12900ks on his Hero and it is an SP84 and a dog. Even at stock (5.2) we couldn't dial it down any lower than ~272w under CB23. 5.2 all core I have mine down to 227w under CB23 running 10min stress test. His 5.3 all core was 319w vs my 255w both on auto. I still need to dial it in to see how low it can go. I jokingly told him I think his chip was meant for me. He wasn't amused. I'm sure he'll play around with it over the next few days and at least shave a little of that off.
  14. electrosoft
    Nice write up by someone with a 5800X3D in hand over on OCN here: https://www.overclock.net/threads/5800x3d-owners.1798046/#post-28975466
  15. electrosoft
    Having used the Noctua LH-N9 series of coolers before, it definitely won't cool a 12900k and I suspect even a 12700k would give it fits. I built a system out for my brother (mATX) and it had his 5600x running a bit toasty even under load. I switched him to an EVGA CLC 240 and that fixed the thermal problems. I guess it depends on what CPU you're looking to run. @Clamibot Linked their new upcoming TREGO pics in the other thread (https://imgur.com/gallery/pT7GUFc) and that seems more my speed being able to get a 280mm or 360mm AIO in there.
  16. electrosoft
    Only problem now is I need to re-certify my 2x8GB (part of my original 4x8GB) set of sticks to see where they will go past 4133 as I never took them that far before. Using them on my Z390, they were 4x8 @ 4000 Using them on my 5800x its IF 1:1 crapped out basically at 3733 My best 11900k crapped out at 3933 My best 12900k was 4133 I'll end up putting them back in the 10th gen test bed (Asus Z590 Prime-A) and getting some known working loose profiles >4133 to make sure they're not the problem and see what happens but for DDR4 I know the real magic is with an MSI Z690 Edge. It is clearly the superior DDR4 board for Z690 now that time has passed and data has trickled in. I see a lot of D4 owners either abandoning their D4s for it or just going for DDR5.
  17. electrosoft
    I'm not even sure how I missed this thread but @FredSRichardson dropped a link in another thread. This is a very cool and fascinating project, awesome!
  18. electrosoft
    I do like the perfect square and thinner width. Makes it easier to just place a portable display right on top and maybe even a TKL Keyboard there too for transport. 360mm AIO basically means hello 12th gen comfortably.
  19. electrosoft
    Yeah, the IMC on this SP94 12900ks is superior too. On my SP91 I topped out at 4000 on 2x16 DR B-Die (4133 2x8 SR B-die). on the 12900ks they booted right up at 4133 and ran just fine. In addition, on the SP91 12900k the 2x16GB DR B-Die sticks could do 4000 but the bandwidth was capped at 58k and latency was 68n+ and they acted odd and laggy. Another reason I still stuck with 2x8GB @ 4000. With the SP94 12900k and these sticks, they are right where they should be pre-tuning.
  20. electrosoft
    Nice ~10% OC on the GPU. That can make the difference when you're trying to avoid those dips.
  21. electrosoft
    I spent awhile (joyously) fine tuning my X170SM-G but now it runs like a beast. I had an X170KM-G but it was problematic and no Prema BIOS for full control. @Clamibot and I have the same grand vision of a killer, well thought out, traveling SFF monster and a portable display and keyboard. I am curious to see where this X15 goes though.
  22. electrosoft
    Doing some auto prelim tests before I crash.... SP94 12900ks same P102 core rating as my SP91 12900k but clearly better: Stock G1 4000 (2x8) Pure auto for both (CB23 multi run): SP91: X142K440...G1 4000 PCore / ECore Pull Watts Temp --------------------------------- 5.2 / 3.9 1.332 262w 82 5.3 / 3.9 1.412 317w 96 SP94: X205L654...G1 4000 (12900KS) PCore / ECore Pull Watts Temp --------------------------------- 5.2 / 4.0 1.288 238w 82 5.3 / 4.0 1.314 255w 87 For kicks, here was my SP82 12900k: X141K007...G1 4000 PCore / ECore Pull Watts Temp --------------------------------- 5.2 / Stock 1.385 297w 94 5.3 / Stock 1.421 336w TT SP89 and SP81 wouldn't boot at all to 5.3 Definitely feels like 11th gen 11900k days again where 5.3 testing saw a high of 370w down to a low of 295w. Then again testing 10th gen 10900k's at 5.2 saw a high of 367w+ down to a low of 261w. All on auto.... Intel, extreme variance is thy name..... I'll see where the IMC lands over the next few days for DDR4 G1. SP81 and SP89 topped out at 3867 SP82 did 4000 SP91 did 4133
  23. electrosoft
    If I had some Kapton on hand I would have used it but I had multiple S33 so that worked fine till I sold it a year later but if given a choice the Kapton.
  24. electrosoft
    I've already committed to my next build or build adjustments being a small custom water loop w/ flexible tubing and QDCs for my CPU at least. We're reaching that point.
  25. electrosoft
    Well.... I stand corrected. 12900ks better than my SP91 12900k SP94 P = 102 E = 79
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. Terms of Use