SteveRo Posted November 16, 2012 Posted November 16, 2012 Thank you stevero Also just another question.. is the WMV Registry tweaking (your encoding score is 3~4 times higher), PowerToy tweak legal ? Yes, powertoy (originally introduced to us by Master Pro) has been legal for sometime.
Toolius Posted November 16, 2012 Posted November 16, 2012 Thank you Im soo sorry for asking all these questions but i dont want to do anything that is not allowed.. thank you for all the help Stevero Guys i have another question... how is 9000 MB on Virus scan possible ?
SteveRo Posted November 16, 2012 Posted November 16, 2012 how is 9000 MB on Virus scan possible ? HDD sector size >512 bytes can get you VS 9000+ which is (soon to be officially) illegal and all submissions using sector size other than 512 should be deleted due to "altering the perceived speed of the benchmark program, tricking it to believe it ran faster ".
Crew pro Posted November 16, 2012 Crew Posted November 16, 2012 understanding a press release will be put together in the next few days i look forward to clarification on the official read on how hwbot see tricking benchmarks into giving a score that’s not real fits with software caching in pcm moving forward. dude i have taken your comments on board, along with every elses, you need to understand that sometimes the rules/decisions arnt just clean cut, this is a very complex decision on something that has long been allowed, in an ideal world we would ban all RST submissions using memory caching but how about while were at it lets remove changing visual quality settings in drivers, LOD, run order in 3dmark01, the 3dmark01 tweak used in z77 bios, as these are all things that change the perception of the benchmark end result, some we would probably remove now if we could, some would stay, plus what about all the users have submitted scores that have been seen as totally legitimate for the past 12-18 months, this is a totally difference issue than these so called "super-tweaks" as are plain and simply cheats, the moderators of hwbot believe this is the fairest and best compromise to go forward, i dont want to hear that talk about how allowing RST is taking harddrives out of the equation, how? virusscan is still the domain of areca/raid controllers, if we have the same setup, we both have dual ssd for virusscan, and i have a areca for my general access? who do think is winning, me and by a reasonable margin if anything allowing RST brings hdd more into play since you need at last 2 marvell ssd, aswell as an areca controller w/ cache and some sort of acard attached to that, its now more storage hungrg
Yamunsa Posted November 16, 2012 Posted November 16, 2012 well it was worth a shot - RST added to the too hard basket :celebration:
Crew pro Posted November 16, 2012 Crew Posted November 16, 2012 well it was worth a shot - RST added to the too hard basket :celebration: put that too hard basket ish away, the mods here put in huge hours giving you a clean/fun environment to bench, if they thought for a second by spending months cleaning up the db they could make things fairer these guys would do it, one of the key issues here is all the guys that have put up submissions under the current standing rules, put up a score, knowing they are within the rules, knowing it wont be taken down in 12/18 months time, this is where things get complicated, this is one of the key reasons rst/raid xpert will stay allowed and only sector size changes will be outlawed, sure some of these guys might have used sector size changes and they will be removed there is one key difference between sector size changes and rst using some ram caching and thats the whole benchmark speed perception issue, ram caching isnt decieving anything, its a real world speed, but its something we have long outlawed due to it out weighting hardware we want to shine throught in these rankings, changing the sector size is decieving pcmark, as its poorly written and not designed to deal with those changes, its the equivalent of changing the basketball hoop from regulation size to 2 foot high it seems we are argueing removing rst just for the sake of it, sure it goes against a long outlined pcmark rule of "no ram disk" but it isnt and hasnt been a problem, life isnt one straight line dude, sometimes different decisions are required that will achieve a similar or even better outcome for me personally and totally isolated from my hwbot decision making, pcmark without rst would be boring now, its a great little twist that adds some compexity into anyones rig wanting to push for high pcmark points
Moose83 Posted November 16, 2012 Posted November 16, 2012 I think this decission is good;) All old scores are safe, and we areca benchers can use RST for nice VS My sector size scores are now off. Mr. Pro, you write again with Team.AU Account:)
Moose83 Posted November 17, 2012 Posted November 17, 2012 So we all talk about what is right and whats wrong...i found an interesting submit, wich pretty shure is outside actually rules... 1,5G VS with sempron singlecore... never possible on 2 acards at SB;) http://hwbot.org/submission/2257400_el_gappo_pcmark_2005_sempron_145_23485_marks
SteveRo Posted November 17, 2012 Posted November 17, 2012 So we all talk about what is right and whats wrong...i found an interesting submit, wich pretty shure is outside actually rules... 1,5G VS with sempron singlecore... never possible on 2 acards at SB;)http://hwbot.org/submission/2257400_el_gappo_pcmark_2005_sempron_145_23485_marks Mr. Moose, yes, agree. Mr. ElGappo, Sir, please delete this submission. I assume this was with (what is now illegal) sector size change?
El Gappo Posted November 17, 2012 Posted November 17, 2012 It would be much higher if it were. Wasn't even available then and I didn't use it when it was. Where are you at with VS and 2/2.5 acards on SB950 Steve?
Moose83 Posted November 17, 2012 Posted November 17, 2012 There are more than 3 guyz, all say its not possible! 1,5-2G VS is possible, but not on singlecore. I wonder how much you can score with X4... 3G? And sector size change is easy to made with all Plattform, no need for raid expert!
knopflerbruce Posted November 17, 2012 Posted November 17, 2012 Are there no similar scores this one can be compared to?
Moose83 Posted November 17, 2012 Posted November 17, 2012 No, he is the only one that have 1,5G VS on sempron;)
Moose83 Posted November 17, 2012 Posted November 17, 2012 (edited) Isnt it enough, that no one can score this? He beat guyz that have 500MHz more CPU Power... SteveRo has 990FX at home, im too.And 1,5G VS is no problem onboard, but only on more than 1 core. Normal range is between 800 and 1G on 990FX with sempron;) Edited November 17, 2012 by Moose83
El Gappo Posted November 17, 2012 Posted November 17, 2012 (edited) There is nobody with 500mhz more on the CPU with onboard raid-0. You should know onboard takes a shit after 338htt or so. Remember a short while ago when you were accusing TD of changing files for having near 10K vs using raid xpert on single core because you couldn't match it? Or last week when you accused him of using flash2desktop because you didn't know vista produced higher TW? Same shit all over again. Edited November 17, 2012 by El Gappo
knopflerbruce Posted November 17, 2012 Posted November 17, 2012 Isnt it enough, that no one can score this?He beat guyz that have 500MHz more CPU Power... SteveRo has 990FX at home, im too.And 1,5G VS is no problem onboard, but only on more than 1 core. Normal range is between 800 and 1G on 990FX with sempron;) So far I haven't seen this score being compared to any specific ones. If it - for an unknown reason - is 50% higher than other scores that can be used for a comparison, then it needs some explaining.
I.M.O.G. Posted November 17, 2012 Posted November 17, 2012 I ran the same setup (same board, onboard raid0), couldn't get anything like those virus scan results. Tried with onboard and areca. My actual subs are done with the best scores I could manage. If you run single core, you know that virus scan ran alone scores way higher than virus scan ran during mtt3 due to the other tests running at the same time. The score isn't reproducible, but I don't care if it gets removed. If everyone wants to keep running pcm05, gonna have to accept that sometimes there are exceptional subtest scores that aren't easily reproducible. If you dont like it, pcm05 isn't a good place to play. There are bigger problems, like top 10 global cpu frequency records on known bugged cpuz versions, where scores were 300mhz higher than submissions made previously on the same chips, and no high frequency sp1m or sp32m from those chips. It's hwbot... Sometimes its hard to know if results are bugged or legit. I see little point in fixating over it, bench as well as you can, and an exceptional score here or there just happens sometimes.
Moose83 Posted November 17, 2012 Posted November 17, 2012 So now there are 4 Guyz that can confirm this VS isnt possible You want more than 50% needs confirme, than this is an case where it is needed. 1000 VS vs. his 1500VS;)
SteveRo Posted November 17, 2012 Posted November 17, 2012 (edited) It would be much higher if it were. Wasn't even available then and I didn't use it when it was. Where are you at with VS and 2/2.5 acards on SB950 Steve? Good afternoon Sir, With Areca 1880/4GB 15xR0 Acard 9010 when using 1x Sempron my max virus scan score is 300-400, actually I don't think I have ever hit 400 with it. edit - I stayed away from sb950 as it looked to me like raid expert was bugged for pcm05. edit2 - Correction - looks like I did get 408 once Edited November 17, 2012 by SteveRo
Moose83 Posted November 17, 2012 Posted November 17, 2012 Steve around 400 VS is normal for Areca:) On SB with write back normal is 800-1000, 1000 only in an lucky run. But only with Vista VS is good with onboard raid, win7 sucks there;)
crustytheclown Posted November 17, 2012 Posted November 17, 2012 I've managed 1200+ VS on single core Sempron but that was on a LSI controller with just 4 M4's...No way for me to go there with onboard controller...I think the catch is that on Vista that single core utilazitation is different from 7 AFAIK that's why you can see 1000 on the onboard controller...Maybe i will retry that theory in the coming days...
Crew pro Posted November 18, 2012 Crew Posted November 18, 2012 hey guys, please leave the moderating to the moderators, we appreciate your assistance in reporting submissions you dont believe are correct, but please leave it at that, if you think nothing is being done or its slipped between the cracks, please contact an administrator
IanCutress Posted November 18, 2012 Posted November 18, 2012 There are more than 3 guyz, all say its not possible! So now there are 4 Guyz that can confirm this VS isnt possible This is quite positively the worst argument I have ever heard. 'X people said it, so it must be true!' Please Mr Moose, come back with evidence. You yourself continuously state that sometimes VS is high and 'it isn't a bug' but just sometimes you score higher. If that is the case then apply some damn statistics, work out your mean and standard deviations for a distribution curve, and actually calculate the possibility of a score getting so high on one of these 'golden runs'. Things that happen 1 in 64,000,000 times actually happen pretty often. To say otherwise is to severely underestimate the number of things that happen.
Recommended Posts