Jump to content
NotebookTalk

jaybee83

Member
  • Posts

    2,041
  • Joined

  • Last visited

  • Days Won

    40

Everything posted by jaybee83

  1. that sux man, sorry to hear that. but im along the same lines, prefer to offer my stuff here first before trying on ebay. easier and also cheaper to deal with any issues 🙂 sorry for the OT, carry on with the sale! 🙂
  2. oh absolutely, this aint my first rodeo into the RAM tuning foray 😄 its by far the most fincky component to tackle vs. GPU (pretty straightforward and quick) and CPU (medium complex and more time consuming, but now with Zen 4 it gets more complex vs. Intel). also takes the most time to determine true everyday stability. one needs to include long idle and low load tests, as well, and do that over the course of days to weeks. and ure right, my experiences so far are just limited to pre-DDR5, havent done much tuning yet on my AM5 system. hmmmm....i sure hope this was due to the board and has nothing to do with your cpu. best to check the pads on the underside, see if theres anything iffy going on. keep us updated on this bud! so far no issues here on the RAM front, just set the EXPO profile (6600), dialed it down to 6000 and off i went for daily stability. currently still tuning the GPU, so havent gotten around to CPU and RAM yet. figured id start with the most expensive component and then work my way down haha 😛
  3. so far ive always had the experience that manual tuning takes u much further than setting XMP (which most of the times doesnt work as intended lulz). just too many variables in mobo, dimm quality / temps and cpu imc capabilities to be able and expect XMP to just "work" (by XMP i also include EXPO, of course). and yep, id also tune it down step by step and see how far he could get. then take it from the back up manually including higher voltages and maybe also better temps on the DIMMs.
  4. double stock over 4090?! no way... 😄 also, that "id wager" seems just more like a hunch rather than having sources telling him stuff... in any case, its all just SMH, i dont get it...
  5. no matter how much noob or master is in you, ure very welcome here bud! 🙂 enjoy
  6. 95C temp target is totally fine if it just follow the boost algo and not actually thermal throttling like on other chips. btw, weird naming, wth...
  7. huh? wait what? u went through multiple kits and all errored out on the mobo? more details pls. have u tried playing with IMC / DIMM voltages? most of the time, cant just straight up set XMP and go, RAM is always finicky that way...
  8. damn dude, thats a pretty good price! especially with 30 day warranty. what kinda probs did u have on ebay? annoying / low balling offers? not keeping their word and canceling last minute before ending?
  9. LOL, holy crap that went fast. u blink and u miss Receiver hahaha. congratz on the super quick sale bro Fox! sick deal doesnt even come close to describing this. damn, how are mere mortals like us supposed to compete with such complete packages at such pricing? LOL ill be posting my stuff here soon, wanted to do it already but real life keeps getting in the way (moving to new appartment, getting stuff ready for the baby, etc.)
  10. oh shes a beaut! hope u get it working man! awesome to see so much help already on site here and trying to provide support 🙂 keeping an eye on this!
  11. ha, as mentioned before, biding my time on this one. from my own experience and what ive seen online in the meantime, 522.25 is still the one to go with a 4090, all smooth sailing, no crashes or weird stuff happening 🙂
  12. huh, interesting! havent used a mobile quadro before, so not familiar with those. but in the end its basically the same, with one u also control the other since theyre linked 🙂 exactamente! thats basically how i did it haha. pretty simple and straight forward. awesome, good job! btw, what ive noticed on my 980M was dialing back from OC or Stock settings to powersave settings, ill have to apply the batch file multiple times before everything is set correctly. to streamline this, ull notice that i sometimes repeated a set of commands multiples times to have it all neatly executing inside a single batch file.
  13. hm...not sure what u mean? from the reviews ive seen so far, the 4080 mostly stays a bit below 320W, more like 300. similar behaviour as the 4090, which would mostly stay below 530 / 600W, depending on the vBios. just gotta select specific loads to hit the power limit before u get voltage limited. also, the 4080 lacks so many vRAM chips and transistors, might just be overall lower power needs? combine that with identical 4090 coolers so temps are waaay low and add to the lower power consumption. are there any reports for differing boosting behaviour? or am i just missing something here? 😅 haha are u sure thats a wise approach to give Intel Arc GPUs for ur wife and kid? this shouts 24/7 maintenance and troubleshooting to me whenever they run into any issues... or are the drivers stable enough to leave them alone with it? wtf, i just dont get it... seriously. and with every day passing, boy am i glad i got a 4090 and at MSRP at that on launch day. couldnt be happier! maybe there might be some middle ground, i.e. specific to private sellers theyd need to hold onto the card for at least a month or so before reselling. similar laws apply to i.e. stock investments, with many countries in the EU forcing u to pay "speculation tax" if u dont keep said stocks long enough and sell them off right away. thus, long term investments are encouraged vs. short term gambling runs. hmmm at just 0.05 to 0.1%, with over 100,000 4090s sold already, were talking roughly 50-100+ cases. so basically the ones posted on reddit and other forums.... that seems awfully close to one another, what a "coincidence" 😛 and so it begins... 😛 one conclusion ive come to over the years is that a lot of ppl hold out on feedback on ebay just to be on the safe side when it comes to "revenge" feedback. i.e. if u give them negative feedback they strike back likewise, no matter if you actually did smth wrong or not. thus forcing u to think twice if u wanna leave any critical feedback. just sad how ppl think sometimes... not sure this is an indicator yet for the 4080s "popularity", rumor mill so far says that the 4080 stock on launch was waaaaaaaay below 4090 levels and sometimes non-existent in some shops. so with just a handful of cards around ull always find some idiots grabbing those to be "first" 😄 also cuz its "new, shiny" and, of course, always better (= new!) 😛
  14. if you go to the overclocking section in Nvidia Inspector, ull see that u can only adjust GPU core and vRAM. there is no separate slider for Shaders. I guess those would be adjusted automatically with the GPU core at a ratio of 2:1 like listed in your screenshot.
  15. u forget one very important thing in these calculations: the machine cannot be THIN anymore! and, as we all know, this is the absolute most important aspect of all mobile machines 😛
  16. yep, thats the same version i got for my lady's machine. i still think its weird why they changed the design to stock IHS format, one of their main selling points WAS actually the increased thermal mass and surface area vs. stock. had a good laugh at that post, so much back and forth and drama with that EVGA OC type and now ure just selling it off to a stranger for 200 bucks more, no hassle, no drama. good for you man. my prediction: now that ure spoiled with the 13900KF golden sample, ull be totally disappointed with the next one u get 😄
  17. alrighty, back from the honeymoon and a bit of time to help out my fellow forum mates 🙂 I have attached the batch files I was using previously with my 980M, generated the commands using Nvidia Inspector shortcuts (the commands are listed in the shortcuts details when right clicking on their properties). as for autostarting, there are three different options that ive used previously: 1) Regedit go to: hkey_local_machine / software / microsoft / windows / current version / run and add a string value with the location of the batch file to run, i.e. C:\Files\Name.Bat 2) Regedit As above, only use a different location for the string value: hkey_local_machine\software\wow6432node\microsoft\windows\currentversion\run 3) Task Scheduler Create Task - General / Add Name / Run with highest priviliges / Windows 10 Triggers - New / At Log On / (Optionally Delay Task for 15-20 seconds if any issue applying) Actions / New / Start a Program / Input Program or File Location (in our case the location of the batch file) hope this helps! cheers PS: Notice how the three batch files have sometimes differing command structures, this depends on what kinda settings u applied in Nvidia Inspector, be it underclocking / power limits or overclocking. Nvidia 980M Inspector Batch Files.zip
  18. ha, im not surprised that smth like this would some from a highly technical vendor like Noctua tbh 😄 and yes, in theory they might as well just throw out their own W/mK value, but im sure theyre aware that customers would be too easily drawn in and just go by that number alone. as we all know by now, that doesnt tell the whole story though. so its tricky going either way, do you let the product speak for itself and risk making customers think u have smth to hide or do u throw out some quick numbers to try and draw in lazy customers? um....wtf? can u contact them and try for a refund or replacement?
  19. say whaaaa...? who does that? literally NOBODY i know, especially not around here 🤣 oh damn, that seemed super fast! i guess theres pros to both approaches, since it would be a waste to just throw away every item coming in and replace it with a new one. but i could imagine that EVGA tries to repair those items in the background while sending out a fully tested and functional one to avoid making the customer wait too long. Asus seems to be going the "1 for 1" route, where the customer gets EXACTLY the same item back unless its absolutely not repairable. speaking of Asus, still waiting on the cashback for my X670E board, lets see when that arrives... been roughly a month now 😅 so first thing i did when coming back from our honeymoon was open up Hyperion and check the gpu power cable 😄 all good though, not even a slight sign of burns / parts melting. also, i liked how super difficult it actually was to unplug the thing, it was 101% fully seated and connected! but definitely will not be doing that checking regularly, dont want to actually cause the thing im trying to avoid lol its fine that they dont update with more details yet, id prefer a thorough investigation with definitive answers rather than 50 updates with half-assed guesswork (as is currently going on with a lots of parties). hahaha i was also seeing all blue tbh 😄 either way tho, both nice colors 😉 yeah also saw that vid, seems to me that "third party" definition is kinda complex, seeing as CableMod actually worked WITH NVIDIA DIRECTLY in designing their right angle connectors 😄 i mean, in general this makes sense, i wouldnt wanna warranty my 2k$ product when its being used with 1$ trash accessories from AliExpress and the likes...
  20. oh for sure, AM5 desktop socket in a laptop would be awesome. especially cuz with the 95C temp target the cpu could just boost all the way until the limit of the cooling solution, so with each new cooling tweak ull get a bit more performance automatically 🙂 issue i see here is similar to AM4 DTR laptops previously: even though physically and electrically, all AM4 cpus DO fit, not all are compatible due to missing bios support. and THAT sux even more than having soldered components, because theres basically NO reason you couldnt upgrade except manufacturer being lazy or just baiting u with upgradeability but then switching / pushing u to buy new prematurely anyways...
  21. hmm not really sure about that @cheaper to have own fab & packaging. reason why Intel is considering to offer their process nodes to external customers in the future (like TSMC) is exactly because having their own fab and doing additional R&D for node advancement is so friggin expensive. on the plus side, u can customize the process fully to fit your design instead of having to make compromises when outsourcing the manufacturing like AMD or Nvidia. and, of course, u always get fab time because its your own fab 😄 so u dont have to compete with other customers on manuf time. id say its a give and take, no real overall winner here. just have to make the strategy work for your products. depends how u look at it. for intel, its 2 years and ure done, with RPL specifically ure EOL right away regarding the platform. for AM5 its currently "2025+" so were talking minimum 3 years, likely longer. of course we cannot foresee what kinda products will launch during the whole AM5 lifetime, so naturally its a bit of a gamble. but following your kind of argument is dangerous: because then we might as well just go all BGA or stick to one platform + 1 cpu gen and thats it. do you want choice and longer platform support or do u want less of both? id go the former any time, any day! 🙂
  22. lulz, just another strat right out of Intel's playbook my man. tell me, how many times have Intel pulled exactly the same thing? now for the first time in ages Intel is finally competing on value cuz they didnt have any other option left. but investors are already angry because they dont like the smaller margins... that should tell you Intel will not be able to keep this up for long https://www.datacenterdynamics.com/en/news/intel-ceo-says-job-cuts-are-on-the-way-as-costs-are-too-high-and-our-margins-are-too-low/
×
×
  • 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