Jump to content
HWBOT Community Forums

yosarianilives

Members
  • Posts

    2266
  • Joined

  • Last visited

  • Days Won

    9

Posts posted by yosarianilives

  1. 2 hours ago, StingerYar said:

    Would be nice & easy for newbies, and pretty comfortable to bench in general, that's for sure. However when it comes to maximum like LN2 - it puts spokes in wheels...

     

    How about some junkie stages, like "Only 6GHz+ submissions", restore "DOGPILE", or maybe do "Slowest GPUPI, or SuperPi"? Sounds like fun :D

    If there is no place for that - how about "only motherboard IGP" stage? Can be fun to tweak & mod.

    I like the idea of Gma, but honestly I think cc should stay ddr3+

    6+ would be cool but I think being able to sub on ambient is nice. Slowest score is interesting 

    Dog pile I'd enjoy cause always nice to bring out collection but again idk about in the comp... 

  2. Well I finally stopped procrastinating and organized the thing. Anyways doing a free bench meet for anyone who wants to make it over to my neck of the woods, renting a venue and a bunch of ln2. Event is sponsored by Intel and International Computer Concepts (an overclocked server company).

    Cost to attendees: FREE!

    Dates: Nov 17-19 & Dec 15-17

    Location: 1801 Eastland Drive, Bloomington, IL 61704

    Who's invited: Literally everyone! US or not 

    Comment below with comments or questions or just to show interest!

    poster3.thumb.png.5db15723f0a27feb5d7008c15f1073c3.png

     

     

     

    • Like 2
  3. 2 hours ago, Tzk said:

    If two categories are introduced for AL, then what happens with benchmarks which won't scale past 16 Threads (so 8 big cores + SMT)? Could a user just submit these benchmarks twice and get points for both? A good example is 32M, as it doesn't scale at all. Could i submit it to 8+8 and 8+0 category? 

    I also think that having to disable cores to participate in a certain benchmark will cause issues and faulty submissions. So we need to have a ranking for "all cores enabled" to simply have the users submit into the right category without thinking too much. I'm also not sure if a separate mixed category will help us at all...

    So i vote "AL = 16C" and done.

    Like I said single global listing per cpu, so alder would be 8+8 or 16, this whole 8+0 listing would be just a way to cheat. We don't let 11900k participate in 4 core rankings so why let 12900k participate in 8 core. 

    Superpi 32m would be single core as per normal, and if we're lucky wprime 32m will go away ;)

    • Like 1
  4. I mean people should stop acting like the slower cores are shit, they're still supposed to be faster than rkl. So 16 core alder is still way faster than 16 core rkl. If the point of new categories is to make it more fair then let's not do that as it would still be unfair, we never allowed down core to compete in more categories because it would still be unfair. Higher core count cpus have more advantage than just more cores, they have extra cache etc. If Intel wanted it to compete in 8 core they wouldn't call it 16 core. 

     

    The only way aside from just counting all the cores I see as fair is if you make a category called 8+8 that's separate from 8 core and 16 core. Would need to go back and do it for all the big little arm cpus as well so might create a lot of work over time, but anything other than just counting the cores will create a lot of work. 

    Regardless I think each cpu should only count towards one global category per benchmark, so either a 16 core or as 8+8

    • Like 1
  5. 15 hours ago, TerraRaptor said:

    Probably make all stages BenchMate-enabled? I don't really like BM but as it should be our future, let CC2021 be another testdrive for benchmate. I suppose it will limit hardware to 478+/754+ but I don't think many are interested in too old hardware anyway.

    Yes, cc is usually new. Tbh I wouldn't be surprised if there wasn't any stages that ddr4 wasn't allowed. 

  6. Now that TC is over its a good time for the community to start suggesting stages for country cup. Should help keep us occupied while leeg is busy moderating scores. 

    Anyways for my suggestions I think some basics, should be a 32m stage, igp stage, ycruncher stage at minimum. Also should keep it to newer hardware as is tradition for cc

     

    Comment below your thoughts and whatever you're thinking 

  7. 1 hour ago, Sparks.nl said:

    For me it is just that I want to use that what is stated. In those lines is no exception for cpu’s which also run on AM3. 
    There was also no limitation for example for C2 steppings. Of course Vishera was not meant to run here. 

    Besides that I see the same architecture in some AM3 cpu’s. It seems like a strange limitation. I even named the 550BE which is only C2. I admit that it dies not happen that much that a cpu works on 2 platforms… 

    In IT audits these things are essential (part of my daily work). 
     

    edit: I am not sure I fully understand this. 

    If s771 cpus are not s775 then am2+ is not am3, not sure where you're getting confused. Competition is by cpu socket not architecture. If you break the pin off of a 940 be it still doesn't work on am3 motherboard. It's an am2+ cpu. If you put a 550 be into am2+ mobo it's still an am3 cpu. 

  8. 1 hour ago, Sparks.nl said:

    Just for the record, I think I named the 550BE (not 555BE).

    What I miss here is looking at the bigger picture. I can cut that pin if needed.
    What is stated in the competition is: 

    • Only use DDR2 SDRAM memory.
    • Only use processors using socket AM2, AM2+ socket.

    Please note that there was nothing stated here about processors. 

    I think that the rules as stated are a good starting point and are correct. 
    Taking this as a reference (for example) a Phenom II X2 550BE cpu has the exact same architecture as the X4 940BE. I can’t help it that AMD made versions which have both an DDR2 and DDR3 controller and assigned some to socket AM2+ and some to AM3. 
    The only clear difference is that AM3 versions officially also support AM3 and with it DDR3. This advantage can only be used using an AM3 socket and DDR3. When used it in an AM2+ socket, these cpu’s work exactly the same as the AM2+ Phenom II since it has the exact same architecture. So to me this still is by the rules as stated in the competition. 

    Is a socket 771 cpu suddenly socket 775 if you put it in a s775 board? Same architecture. We go off of native socket of the cpu. Am2+ cpus physically don't fit in am3. Also I can say that the am2+ phenom IIs clock so terrible that it almost feels different lol

  9. 1 hour ago, GtiJason said:

    Well it's not like you can just run a Coffee Lake -R cpu with DDR3. There were only a handful of Z170 chipset 1151 boards made and they were 1 not very popular and 2 never meant for anything more than a budget solution to jump to Skylake. Some of the boards only officially supported up to 6700k cpu with 7700k being as far as these boards were designed to safely push.
    So just getting an 8700k or 9900k to post requires modification to the bios as well as patching and finding a way to flash the modified bios without bricking the board .     

    You also need physical modification the pcb itself, more specifically the Nuvoton chip / skt_occ.

    Once you have all that working then you need additional modifications to feed enough vcore to the 8c cpu and you can see more issues like cooking the power delivery planes to a crisp to deal with etc.
    So yeah oversight and loopholes have owned these competitions for years and if you're willing to track down such rare hardware and put all that work and testing involved just to run 1 or 2 benchmarks you deserve to be rewarded and 9900k cores and ipc is said reward. Good work Tagg
     

    Z170 9900k.jpg

    SKTOCC.png

    VMod_Mod.jpg

    Don't have to short anything on board, just gotta short two pads on cpu so it posts on top of modding bios. Tagg just wants extra vdimm lol

  10. I think that's IF we're going to require newer os, maybe require benchmate etc then making it from certain architecture onwards seems like the best way. That way we don't end up with unbeatable legacy score but we can also implement newer requirements. 

  11. 55 minutes ago, TerraRaptor said:

    Looking how great pcmark is in scaling with greater core count, I would doubt it. Both cpus will share just 2 channels of ddr2 (that sould also be a challenge to overclock). 

    Though, I'm curious of the result.

    Imho the biggest handicap is the pcie arrangement, pcm is mostly about storage speed. That's why x48 might still be faster than p45 because 32 pcie lanes vs 16. So you could do a full 16x raid card instead of 8x. All the s771 chipsets have a good amount of lanes but they're all split into groups of 4x and 8x with no way for a 16x slot. Raiding multiple cards isn't gonna happen on a bare metal install so x48 is gonna have fastest storage. 

  12. It's relevant for fx because we're making categories for cpus with different core strengths in different core configs. if fx shares part of its instruction pipeline between two cores, not just cache, then creating a one core per module category is no different from creating a 12900k category for disabling all the "shit cores" (they're Skylake speed, stop acting like they're super slow guys). 

     

    I still think that just adding up all the cores is the best, if Intel says it's 16 core then it's 16 core. This approach was used when people wanted fx to be 4 core with ht historically, the decision was if amd says its 8 cores then it's 8 cores. Of course historical decisions don't have to be future decisions as undeniably some historical decisions are probably not the way to move forward. 

    • Like 1
    • Thanks 3
  13. Ht would become a huge handicap until you get above the non-ht thread counts. So currently 8 thread. Throw away 7700k, throw away 5300g,7350k is useless. It's time for 7640x, 9700k, and all that other shit nobody has spent as much time benching hard. 

    Not sure this is good or bad, but that's mainly how it would shake up rankings. Oh and g470 would become even more irrelevant lmao. At least 1, 3, 5 cores wouldn't really change in the slightest 

×
×
  • Create New...