Jump to content
HWBOT Community Forums

mikeguava

Members
 View Profile  See their activity

  • Posts

    24

  • Joined

  • Last visited

 Content Type 

Posts posted by mikeguava

    Hadrware sharing spoted AGAIN?...

    in Submission & member moderation

    Posted · Edited by mikeguava

    Andre cheating???? Lol -he's waaay too slow to be cheating. OPB -is whole different story - he can be so quick sometimes that you don't know what has hit you. But then this gy is relentless - he just doesn't ever stop. Well then so is Andre - relentless...Taiwan benchers rock!

     

    On a serious note - it's waaay to easy to go on a witchhunt with results - sometimes a bencher simply finds a new tweak, a golden CPU or just the perfect combo. Smoke always clears at the end of the day

     

    On a funny note: how many seconds does a Taiwanese minute have?

    PCMark05 HB Rules,FM Rules and correct submision for ranking.

    in General overclocking

    Posted

    wow - another interesting PCMark05 thread...looks my favorite benchmark is getting ready to get canned.

     

    Well maybe the time has come to kiss PCMark05 goodbye. Fact is - HDD scores are too heavily scoring if SDDs/IRAMs/ACARDs are used. It's getting too confusing what is legit and what is not. To the OP - sorry as cool as your score is - like Gautam tried to relay - 9000mb/s as simply not possible by current hardware alone. A single RAID controler in a perfect setting might!!! hit 1200/mbs, 3 PCI-E theorectically could hit about 3000mb/s - but 9000mb/s simply is not possible without the help of some innovative caching. Whether the caching is done by a driver, software solution or software ram drive is secondary - the problem is that the scoring is not representing real world performance of your actual hardware.

    I am using software ramdrives in my 24/7 settings - they truly do make computing a lot more efficient! But at this point in order to "police" the bench I feel it gets too hard to monitor what is legit and what is not if we want to only benchmark the hardware.

    My 30K Pcmark05 score ?

    in General overclocking

    Posted · Edited by mikeguava

    and why is that a problem ?

     

    I'm not one of those that 'heavily opposed' software ramdisks, I've always been for them even before pcmark04/05 existed as they're used widely in the REAL world in server/linux environments to improve storage based I/O performance ;)

     

    why i'm so pro SSD + MFT ? answer is easy I use it to improve my SSD performance for STORAGE related usage in the REAL world = video encoding scratch disk

     

    4x32GB OCZ SSD Raid 0 + MFT on Highpoint 3520LF PCI-E controller with 256MB cache

    vs

    750GB Samsung SATAII

     

    Configurations for System Clocks:

    4009Mhz @19x211Bclk with 8x mem multiplier 3x1GB HCF0 @844Mhz 9-9-9-24

    • i7 920 with HT no Video cache (SATA HDD source drive) = 4min 50s
    • i7 920 with HT with Video cache (SATA HDD source drive) = 4min 10s
    • i7 920 without HT no Video cache (SATA HDD source drive) = 4min 34s
    • i7 920 without HT with Video cache (SATA HDD source drive) = 3min 53s
    • i7 920 with HT no Video cache (SDD source drive) = 4min 05s
    • i7 920 with HT with Video cache (SDD source drive) = 4min 06s
    • i7 920 without HT no Video cache (SDD source drive) = 4min 04s
    • i7 920 without HT with Video cache (SDD source drive) = 3min 55s

     

    xvid_times_chart.png

     

    Note: Video and Audio caches use system's tri-channel DDR3 memory bandwidth which is faster than both SATAII and SSD Raid 0 transfer bandwidth and as I tested early if you disable HT I notice that your get more clock for clock memory bandwidth in Memtest86+ v2.11 than when you have HT enabled.

     

    Guess ultimately it comes down to definition of storage usage. The fact is with SSD and even software ramdisks (which were out there way before pcmark04/05), the definition of storage has changed. Now the question is whether FM/hwbot accepts this evolution or not.

     

    Anyway, it's up to you guys to decide whether MFT assisted SSD is allowed or not. It's your house so it's your rules either way.

     

     

    edit: Also if the argument is pcmark05 is to test hdd performance not system memory performance which software ramdisks can effect, then the argument can be the same for 3dmark2001/2003/05/06 which all are effected by system memory bandwidth to various extents. Classic is 3dmark2001 which can get pretty big boosts when memory bandwidth is tweaked. Accordingly, since 3dmark benchmarks are meant to test GPU performance and not system memory bandwidth, we should ban 3dmark2001/2003/05/06 as well.

     

     

    So in this example it seems like that using the regular HDD with HT disabled is better then SSD+MFT ???

     

    The problem I personally have is not the real world performance gains in applications, and again MFT is great for SSD and all other drives as well. But as you can the the real world speed improvements are minor and are not 20X the performance which we see in HDD tests in PCMark05.

     

    For the way PCMark05 calculates scores allowing softwares RAMdrives and MFT tweaks the total socres get completely messed up. The 220MB limit kept the score skewing somewhat under check - a fast quad core system will outscore a dual core system. With software ram scores the entire benchmark score system gets ruined and whoever runs the best memory will get the top scores. Isn't there SuperPI etc for that?

     

    I tried to make the point by submitting my untweaked AMD air benched score which kills a LN2 bench of mine with Core i7 @ 5GHZ. Allowing HDD based score to get that crazy kills PCMark05 in my point of view.

    I just ran my AMD system with an IDE drive - I got 12k, with MFT I get 35K - 23000points coming from HHD based points is just way too far off.

    My 30K Pcmark05 score ?

    in General overclocking

    Posted

    From easyco install manual:

     

    From easyco install manual:

     

     

    Running MFT Performance Benchmarks

     

     

    You can run standard benchmarks against MFT and you will be astounded at some of the results you get.

    That said, some may overstate MFT’s performance because MFT’s performance is, in principal, dependent

    upon the degree of free space, and we increase free space by virtualizing sectors expressed as all zeros.

    Thus, some tests may indicate that we are performing IO at gigabytes per second when the same is physically impossible.

     

     

     

     

     

     

     

     

     

    MFT runs nice nevertheless - beats some of my software ramdrives.

    Allowing MFT and or SOftware ram will allow more peeps to competively bench PCMark05 - thsi at least would be a bonus. Just the idea of an overall system bench will get lost

    My 30K Pcmark05 score ?

    in General overclocking

    Posted

    Yeah it seems like MFT is a FIFO queue...you write to this software queue first, then it writes to the actual array in a linear fashion. So it's very much a software buffer explaining why it performs like a software ramdrive.

     

    Well...if MFT is to be allowed then your score should be as well. :P: If not then neither should be.

     

    I don't want my score to be allowed - that was sooo slow ...

    My 30K Pcmark05 score ?

    in General overclocking

    Posted · Edited by mikeguava

    I second the notion that if 05 is removed I think vantage should get points. I really enjoy the pcmark benching. Its not as popular, but there is alot of tweaking involved.

     

    Vantage has an advantage also being that the drive you install the OS to is heavily weighted in the score (even if you set a different target drive). So that rules out software based ram drives altogether.

     

    I think now that 3dmark vantage has been approved for points there should be no arguing about pcmark vantage. I.E. the Vista only argument, having to purchase it, etc. The argument that it takes too long to complete is also moot, since people didn't complain when 32M was taking 10+ minutes.

     

    let's just ADD PCMark Vantage and keep PCMark05 :-)

     

    MFT caches software writes operations and send them sequentially to the hardware; it's the software required for the bargain SSD out there to work optimally; in short: it's a ram-like drive indeed :)

    http://www.ocztechnologyforum.com/forum/showthread.php?t=47183

     

    nicely put!!!

    My 30K Pcmark05 score ?

    in General overclocking

    Posted

    The reason I enjoy PCMark the most is that your ENTIRE system gets benched and not just part of it. Obviously there are some problems with it e.g. 3D part does not credit multi gpu etc. but still the PCMarks are great tools to give credit to your entire system.

     

    Again in respect to MFT - I think what MFT does is to setup a software ram drive to store the data temporarly and afterwards to dump it to the HDDs ( whether IRAM , SSD etc. )

    During the PCMark bench the data that is on the software ram drive only gets benched which gives us these artificially high scores.

     

     

     

    Not sure where I claimed to have run software ram in my bench, but not that I care to have it removed - MFT and Software rams is

    the same to me and all should be not be allowed

     

    Kingpin posted a nice run a couple of days ago still maintaining the old 220mb limit - NICE!

    My 30K Pcmark05 score ?

    in General overclocking

    Posted · Edited by mikeguava

    I think that the 225 XP Start up was well done by FM........Now look at you!..... :(

    We now all can get a high score with SSDs or with software ram....Who could tell/find the difference?..

     

    That's all BULLS...... :(

     

    i second that -

     

    but I am pretty sure that there is no difference between software and MFT because I suspect that all that MFT does is to create a software ram to temporarly store the data and later dump the data to the drives.

    MFT works on any drives - not just SSDs

     

    I am all for finding ways to speed up drives - I have been enjoying the game of fast arrays for a long time - but it is sad what has happened to my favorite game.

    My 30K Pcmark05 score ?

    in General overclocking

    Posted · Edited by mikeguava

    wtf...MFT is software, too.

     

     

    Look at this link.

     

    http://downloads.managedflash.com/documentation/090107_windowsinstall.pdf

     

    See chispy's HDD - General Usage score 19xx. It's using MFT software.

     

     

    Why does hwbot accept SSD+MFT score?

     

     

    OK, I will use 4*Acard+MFT break 40K+. You cannot remove my score because you already accept chispy's score.

     

     

    40K sounds low??? check out the dude that had 26000mb/s in Virus check - maybe 50k???

    My 30K Pcmark05 score ?

    in General overclocking

    Posted

    He used Software Based Ramdisk , I used 4x 30GB OCZ Core Series V2 SSds in Raid0 on Adaptec Raid 2405 Raid Controller card ;)

    How can you tell what I used??

     

    MFT + IRAM gives ya about 3.5k XP startup - without fully tweaking. I am confident I would easily break 4k if I wanted.

    Sorry CHispy - tweaks like that have been around for a long time - you and EVA are not the first to find it.

     

    Problem is you are not measuring the HDD speed with an array configured the way you have but memory speed.

     

    Things were fine before this thread started...

    My 30K Pcmark05 score ?

    in General overclocking

    Posted

     

     

    Is the difference noticeable in other harddisk subtests? Haven't been able to test iRam versus software ramdisk.

     

     

     

    You can not tell between software ram and MFT or similar configed drives. But then running MFT on cache drive is not much different from running software ram - data gets dumped without getting written as fast as your system memory can...

     

     

     

     

     

     

     

    At first sight, this certainly seems to be the most elegant solution, but it'll give problems for those who, for instance, want to run Vista using 1 iRam. Plus, like you said, there might be a workaround for this as well.

     

    Works fine in PCMark Vantage to limit ORB benching on OS drive.

    Vista on 1 Iram??? Been doing that for quiet a while...vlite is our friend in need... who needs a Vista that is bigger than 2gb???

    My 30K Pcmark05 score ?

    in General overclocking

    Posted · Edited by mikeguava

    Let's not overreact :).

     

    First of all, even if FM would be allowing software based ramdisks (or just provide verification links for it), HWBot will not accept software ramdisk scores, even WITH verification link.

     

    I reckon it's important to request yet another fix (solution of Mike seems to be the easiest)

     

    the problem will be to distinguish between software ramdrives and creative RAID array solutions etc. For example I can configure a single Iram to hit over 800mb/s at XP startup - pretty much the same 800mb/s I can get with one of the popular free software ramdrives. Now how can we tell if a certain score is an Advanced Hardware Based Drive or a slow software ramdrive?

     

    I personally got into PCMark a long time ago, for the reason of finding the max possible overall systemspeed - which always entailed insane drive arrays. It always is fun for me to find a new way to speed up the transfers, startup etc. I would greatly miss this part of PCMark.

     

    Only way I can see to prevent this as mentioned before - would be only allowing the OS partition to be benchable.

    But even for this solution, there "might" be workarounds - depending on how good the "update - fix" would be coded.

     

    I assume that such an update might require a bit of work from Futuremark's coders - not sure if they have the budget for such a thing at the moment...considering the fact that they haven't been doing anything lately to fix?/upgrade the ORB....

    My 30K Pcmark05 score ?

    in General overclocking

    Posted · Edited by mikeguava

    Ugh, you can't be serious. :-/

     

    Optimally we'd need the limit lifted and constant vigilance from the community. (On hwbot this doesn't tend to be an issue)

     

    problem is - very very hard to distinguish between hardware and software in some cases. I can beat many software ramdrives with hardware - but not all. One way would be to limit PCMark05 only to be benchable/publishable on the OS install drive. This would prevent most software ramdrive based scores...

    Would be a simple update to PCMark05, but doubt that Futuremark has the brains to implement such a fix.

×
×
  • Create New...