Top 20 systems
I've been looking at the top 20 best systems and i can't understand given the top 5 ones , given the cpus, ram, mobo and ssd/hdd how someone got 13117, when the next best one is at 9000 and is actually using a better cpu, currently the fastest cpu there is apart from 10 core E7 cpu .So given windows 7 as a starting point, a mobo with two e5 2687w or a bit slower 2690 should be the fastest scores, even slower are the 3930k and 3960x, yet this guy get 13000 instead of around 9000 ?? And there's nothing in his configuration that could explain it. Unless he was running msdos or something. )
Only thing that's different and higher compared to the e5 2687w systems is the bus speed. Now all those systems have 100 mhz buss, and top at 8000-9000.
Curiously his baseline doesnt show his storage ...the ssd. The only thing that could explain such a massive difference is if he's got windows 7 booting of a ocz z-drive r4 ...lol.
Last edited by Choice777; 06-06-2012 at 11:02 PM.
If you just look at the CPU score, then the machine in 2nd place was faster than the current top machine. But the 2nd machine had rather so so disk performance.
The highest score isn't the highest only because of the CPU. The results are very high across the board. The disk and CD results look like they are both from RAM drives or very high end RAID or PCI solutions. There is a massive 98GB of RAM in the box, the 3D results are good, while the 2D results are extraordinarily high. The CPUs in the machine have a total of 24 cores with hyperthreading.
Still something doesn't add up because it's the same person who has the top 2 score.
13117 2*X5690 @ 3463.4mhz HT on, 98295 MBram, 1486GB NoModel, GTX580 1.5GB 1680*1050
9840 2*X5690 @ 3458mhz HT off, 98286 ram, 298GB No Model, GTX590 1.5GB 800*600
So first 2 are same guy "chrisdildilio" disables HT, adds a little better gpu and gets 36% increase ??? 9840 + 34% is 13185
3rd best score is actually a better set of CPUs and more ram and again gtx590 1.5 gb
9000 2*e5 2687w HT on, 131048 ram, 480 GB ssd, gtx590 1.5gb
So how does he get 13117 when a similar pc gets 9840 ..
The only scores hes got a win on are the 2d
And his 2d score with a gtx580 is 45-50% better than his 2d score with his gtx 590 ??
Also 2d windows interface score is around 800 points with the gtx 590 and 580, while everybody else score about 110-180 with gtx 580/580, ati6950?
Again 2d image rendering score, he gets 7930 points , everyone else gtx580/590/680/ati6950 gets around 600-1000.
On the 3d tests everything returns to normal. Where is his advantage in the 3d ? its missing !???!!
On memory mark , again everything normal, he doesn't score highest in any test, just average to all the top 7-8-9000 points PCs.
In disk test again average compared ot all the 7-8-9000 points pc, but everyone gets owned badly by a 3930k with vertex3 ssd which gets like on average 6-7 times more.
Cd test again, gets owned by other systems.
Basically he gets better 2d score on some tests, even against better cards 580 better than 590 and 680 ? right...
I am not sure what you are referring to. In the top 20 benchmark chart the number #2 result is a E5-2687W CPU with a score of 9000.
Where are you looking?
Yes, as I noted above, the #1 result has 2D results which are extraordinarily high. I had a closer look and decided that a couple of the 2D results are probably faked / manipulated / wrong. So I have deleted the the #1 result and a few other results (from the 600,000 in the database) that exhibit the same issue.
I downloaded the baseline results in the desktop program and compared the ones with the names
chrisdildilio sr-2 dual xeon x5690 .... BL546473 with 13117 points and
chrisdildilio z800 same cpu............. BL392416 with 9840 points and
BOXX technologies.....e5 2687w........ BL601238 with 9000 points.
And 3-4 more systems that are like next best ones with 7-8000 points just to see how much the difference is between gtx 580, 590,680 and ati 6950.
Last edited by Choice777; 06-07-2012 at 10:24 PM.
its simple, just look at how passmark calculates the final score, 21% is from the hd speed, the guy in the first place is using supercache or something like that, is a low level driver in top of the OS that uses the RAM as a disk cache transparently, i'm personally think that this should be fixed as this table has no meaning working that way...
There are half a dozen options available to have super high speed 'disk' storage.
Various RAM drives, Various high end PCI cards, RAID, etc..
Doesn't make sense to ban them all.
(We have also seen cases of the people wanting all hardware that is better than they can afford banned from the charts).
one thing is better hardware and another things is a difference score of 40000000 on the same disk just for using a disk caching in ram, the soft is not benchmarking the disk at all, having all the values separated on a database should be easy to make customs top 20, for specific combination of hardware, removing the hard disk for example...
Originally Posted by David (PassMark)
Not sure where you got this number from. Not sure what a "difference score" is either.
a difference score of 40000000 on the same disk just for using a disk caching in ram
The fact is that solid state drives, RAM drives, hybrid drives & PCI based drives are now fairly common. None of them contain an actual disk (well a hybrid disk might, but it might not be used).
We just can't ban (or even identify) everything that doesn't have a spinning disk in it. And what about drives that do have a disk, but also have additional caching e.g. on a high end RAID controller?
It should also be noted that anything behind a RAID controller (and many SCSI controllers) can't be identified. That is to say, you can't get the model number of the drive via software. Details are hidden behind the controller.
So the division between solid state and rotating disks isn't clear cut.
the guy in the first place of the top 20, CD SCORE = 240270, DISK SCORE = 63267
240k in the CDROM? 63k on disk, that's 26% of the final score, the rest of the components scores are normal, if you cannot detect who is cheating the score then remove the disk and the cdrom from the formula to calculate the rating or at least give less %, i would prefer a separate top 20 list without taking in consideration those values...