wutske

The life we all dream of 😄

I don't think any of the CPU's I've mentioned are any special. It's true that the supply side of AM3 CPU's in Belgium/Netherlands is small, but I can't imagine the demand side being that big either. The only thing I can imagine is that AM3 still officially supported Windows XP, retro gamers like to throw money at that old stuff . Though apparently AGP is more desired than PCIe. The 4770K ... it's a K-sku, so that's the only special thing about it. It also seems to be a professional seller, so there's no info whether it's a golden sample or not. I sure hope CPU's won't become like cars. Can't imagine someone waking up, walking to their wall of CPU's wondering which one they'll be putting in their computer that day 😎

Check your submission id's. The 2nd on is lower (so it's the oldest/first submitted), so chances are it's processed first, then it gets awarded points and then your second submission is processed and since it's better, it means the first one is no longer worth any points. I have no idea how the backend works, but it fees a lot like a it's a queue. If you submit a 100 results you'll see the scores pop-up one by one over a period of time because it's not processing your 100 submissions in parallel.

48 points not enough ?

So, this is something that I've been wondering, especially considering there's people here with 200+ CPU's in their hardware list, so plenty people that probably have experience in this market ? . Recently I got a pretty okay deal on an Asus Crosshair, FX-6200, some mediocre RAM and and a Noctua NH-D14, so now I have a great AM3 board and I'm looking for more CPU's to put into it. Though, when I check the (semi) local pricing for 2nd hand AM3 CPU's I get: * FX-6300: €20 * Phenom II X4 810: €30 * FX-8210: €25 It's already weird that the price goes down if the CPU gets faster, but is it me or are these pricing kinda high for these 12 year old CPU's that weren't even great to begin with ? Intel's 2nd and 3rd gen cpu's were much faster and those also go on sale for about ~€20-€30 nowadays. The next Intel socket goes for €20-€50 with someone asking €75 for a 4770K, which imho is ridiculous because a brand new €75 Ryzen 3 4100 easily beats this CPU. For €10 extra I have a new Ryzen 5 4500 which is almost twice as fast as a 4770K. So now I'm wondering; are 2nd hand CPU prices way too high or have they always been like this ? I can understand that for AM4 the prices are still high since it's still a semi-current generation considering AMD keeps releasing CPU's for this socket, but AM3 is over 10 years old and AM4 replaced it 8 years ago with AM4. The same applies to Intel, why ask €50 for a 10 year old CPU which is 10 "generations" behind the current gen and is 4 sockets behind the current one ? Am I just being cheap or do you have the same feeling about it ? FYI, €1 = $1,09 ... so € is pretty much interchangeable with $.

I don't get it, nobody said you should crank the VSoc, Buildzoid even has his a 1,15V and he even says that high VSoc harms infinity fabric speeds. You can't even change the VCore on X3D chips other than what PBO allows you to do and even then it's within very strict boundaries. You can still crank VDDR whether you have an X3D chip or not. I think a lot of information is lost in translation. You have a good motherboard for overclocking. You have good RAM for overclocking. You have a CPU which is absolutely not made for overclocking and the CPU prevents you from doing that as much as possible for a very good reason. You have to learn how to walk before you can run and with that CPU you're trying to do the 100m sprint with indoor sandals. We are really trying to help you but I personally think you're making it too hard for yourself. There have to be Japanese forums where they can help you without the language barrier. It will also be easier if you have a CPU that you can actually overclock without PBO. Disabling everything that changes the frequencies and power states (PBO, Power management, S3, C1, C6, Cool&Quiet, Turbo, ...) makes overclocking a lot easier. //edit: I checked you hardware library, seems like you have plenty of experience with unlocked CPU's and some nice results too. You have a Ryzen 7700 which did great on this board so I don't get why you're saying it's a bad board. You know what your 7700 can do on that board, so put that one back in and try to optimize that one first using PBO to get a hang of it. Then try ECLK overclocking that CPU instead of using the multiplier and see how far you can get it.

Zen4 can do DDR-8000 but there's this one setting you have to change so that it no longer runs 1:1... If only my brain could remember what it was again ?. Anyhow, DDR5-6000 is still assumed to be the sweetspot for AMD, higher frequencies won't benefit you a lot, it's better to focus on tighter timings. Not being able to run 1T might also be your RAM. I have a set of Kingston HyperX Black that does 1T fine at 1866Mhz. Same board, same CPU but some Geil dragon something memory and it won't even do stock at 2T at 1333Mhz, even though these sticks can do exactly the same RAM timings as the Kingston memory at 1866Mhz, but at 2T. You mentioned you can't overclock the RAM since you've switched to the X3D, so I assume you ran it faster/tighter on another platform ? Your DDR5 kit has Hynix M-die and manufacturers already push a lot out of it using the EXPO profile. If you want it to run faster, either up the voltage or drop the CL. Check out Buildzoids video for more info on tight DDR6000 on AMD: How you find out which core is stronger all depends on your preferences. You can use https://github.com/sp00n/corecycler to find out which core is unstable and then change the offsets for that core until it is stable. You can also do it by hand using y-cruncher. In the component stress tester hit "1", disable all cores with "d" and then enter a single core id (starts from 0). Pick your favorite tests and let it rip If it crashes, update PBO settings and repeat. If it's stable for a couple of hours, then move on to the next core. Repeat until you've done each core and now you'll have a per-core PBO tuning. Then repeat the test with 2, 3, ... active cores and if this thing doesn't crash after hours of running then you should have a 100% stable OC.

You can download it here: http://numberworld.org/y-cruncher/#Download (and it's also included in benchmate). If you run it, just enter the component stress test. Only enabling VST/VT3 should be enough and then let it rip. This will give you and idea of the all core stability. You can change the number of threads to test stuff like single/dual/... core workloads. You're really going for the toughest CPU to overclock and you're doing everything at the same time. Try to optimize the CPU first with PBO only. Once you've got that setting, go for optimizing each core so you get to know which are the stronger and which are the weaker cores.

That would be awesome. I'll see if I can do some better benches this weekend to get a better understanding of what impacts what.

Did you just copy the exact settings from SkatterBencher ? Because there is literally 0 guarantee that that will work for your CPU. Also try testing for stability with y-cruncher, it can really quickly find stability issues which you then have to try to fix by changing PBO settings. It can also be that you've lost the silicon lotterly.

To be honest, SuperPi was surprisingly consistent, all results are within 0,11% ... though that's the problem I'm having and I have seen wprime scale with memory frequency, but that might have also been more of a lucky shot. These are my full timings: I know I need better RAM, the best stick I have is a single stick of 2GB OCZ Platinum which does DDR3-1600 at CL7-7-7 (at 1,9V that is). The Kingston sticks I'm using in this test can do DDR3-1866 at CL10-10-10 at 1,8V...

So, I was trying to figure out which of my DDR3 kits performed best (lowest latency, highest frequency) and I noticed that pretty much nothing I did improved my scores. I used my FX-6200 for this, mainly because it was what was sitting next to me. I overclocked the CPU tot 4,4Ghz, nothing fancy but I needed something stable enough so that crashes could be nothing but RAM. I tested 7-zip, SuperPi 1M and WPrime32 since those benefit more from faster RAM compared to e.g. Cinebench (and they are rather quick to finish). Running DDR3-1833 CL8-8-8, DDR3-1833 CL10-10-10 and DDR3-2133 CL10-11-11 all yielded pretty much the same results, with less than 1% difference. I've read that Bulldozer wasn't exactly the best thing in the world, but ... is it normal to have such little speed difference ? FYI, I was lazy, so for 7-zip green is the best result and for the other red is the best result.

Hey guys, this one seems to be stuck in the processing queue: https://hwbot.org/submission/5442826_ I tried editing the result hoping it would unstuck it, but this didn't help.

Pretty sure I just copied a hwbot submission to pre-fill all the hardware fields. Really have to remember not to do this.

That dropdown is disabled right now.

?‍♂️ Oops, I did it again https://hwbot.org/submission/5442851_wutske_geekbench3___memory_single_core_score_ddr3_sdram_2143_points

Borked this one up, please delete https://hwbot.org/submission/5442831_wutske_reference_frequency_rog_crosshair_v_formula_1_mhz?recalculate=true

I know, I just assumed that I'm in some queue waiting to be moved from league, with all the changes and recalculations going on lately. Anyway, it's a new year, fresh start, as soon as I get my next batch of DI I'll see if I still need it to be fixed.

I think you can always just ask to be put back to enthousiast, but that's more something for the support forum ? It's kind of weird how it works because I did a couple of dry ice subs and I'm still not upgraded to apprentice ?‍♂️

Definately broken om W7. You can also just copy paste the results folder to another system which is up to date

Aight, so I didn't had to get my driver validated after all ?. Well, that's one more valid driver people can use.

Congrats to our down under fellow overclockers ?‍♀️ That's sad unless they find something that's really off... Talking about that, how strict are the rules on valid 3DM results for the HD-graphics stage ?

These may go https://hwbot.org/submission/5420821_wutske_superpi___1m_core_2_duo_e6540_15sec_250ms https://hwbot.org/submission/5418537_wutske_pifast_core_2_duo_e6540_28sec_360ms https://hwbot.org/submission/5421859_wutske_wprime___1024m_core_2_duo_e6540_12min_9sec_851ms //edit: please also delete this one, I seem to have forgotten CPU-z: https://hwbot.org/submission/5119956_wutske_cinebench___r15_core_i7_8650u_723_cb

What exactly doesn't work ?

How do we prevent this from happening again?