K404

Yea, i've kinda thought the same. Aim for gold, or silver, or may as well stick to water for 20-22 points. It's a hard line to walk though. Personally, I feel the point exponent is too steep, but popular No.1 spots ARE impressive.....

Quite mental Congrats!

3:5 ratio = bugged = FAIL?

I'll use the nVidia 8800 series coz.... I know it If it's split up at a simpler level, using the total submissions for the card as the popularity guage.... BUT.... less submitted-to benches get fewer points.... 8800GS (384 & 768): 692 + 23 8800GT 1GB: 523 8800GT 256MB: 446 8800GT 512MB: 8802 8800GTS 320MB: 6546 8800GTS 512MB: 7239 8800GTS 640MB: 6601 8800GTX: 13606 8800 Ultra: 4258 8800GTX gets 50 points per popular gold, but Unigine gets e.g.... 4 points because it has not many submissions. Next card in popularity is the 8800GT 512MB. Thats gets 65% of the points of the 8800GTX... so 32.5 Next.... the GTS 512MB: 53% of the points of the GTX..... 26.5 per popular gold etc etc Least popular card: the 8800GT 256MB: gets around 3.4% of the points OR maybe there is a way of using the total number of submissions: 48736? I suggest that as it stands, every single piece of hardware can get 50 points and nothing automatically stands out as being popular or good

Yep.... fewer big rankings Basically.... my thinking is that once the leaderboard is "saturated," it doesn't matter if there are 300 scores submitted, or 1000 or...... 10000- it's still 49.9 points. I would like to see fewer categories get maximum points, because (IMO,) it doesn't reflect the achievement for the guys taking gold in the MOST popular categories. Ideas: Only have maximum points for the single most popular piece of hardware in a hardware class: as broad as s478/775/1366? Or within a CPU series? Wolfdale, Nehalem? For GPU..... ATI 3 series? nV 8 series? It would mean the popularity is relative. It would also mean that the most-rewarded hardware can change, which I think is a neat idea.... it would (at least partly) reflect the buying patterns of computer users Also, could the definition of most popular be a function of HWB userbase? At the moment, if the threshold is a fixed number, it doesn't "mean" anything. If the saturation point is 300... does that mean there is a user base of 301 and the component is OMG popular? Or the userbase is 30,000 and 1% use it? It's also "fire and forget" which means no announcements stating that some people will lose points because the algorithm is being explicitly changed... and it means the numbers have more meaning to people who do not have a massive back story on HWB, because only HWB users understand the points system. Vaguely related idea....... Should hardware categories be "closed" after a certain length of time? I am on the fence, but it's an idea i'm throwing out there

Yes, a major European (international) party!

2500 No.1 spots is mental. That's saying each of you has benched MINIMUM 250 items with gold in every single benchmark. Respect! Would it be possible to have GPU achievements similar to the CPU socket ones? ....and would it be possible to mask off achievements a user does not want to work on? I have a scrolling page of....130? achievements and most of them I will never try and get, but I have to scroll through them anyway....

Hyperhorn... you're right. I tried to change the thread title just after I posted, but I couldn't/can't. EDIT: Thanks

Ok Marc... a challenge. http://hwbot.org/submission/1025377_turrican_superpi_486sx_25mhz_2days_10h_47min_30sec_0ms How many times can you change your socks in: 2 days, 10 hours, 47 minutes. Keep a count. Dedicate yourself 100% to this. Do not visit HWBot for that length of time

Have you seen the age/performance of some of the CPUs Karl benches? If he cannot change his socks hundreds (thousnads?) of times by the time 1M finishes, maybe he has mobility issues

I am disturbed by the presence of another UK overclocking Kenny

I thought that if two benchers were not on the same team... and the scores were obviously not done at the same time, it was ok? If the same card cannot be used by more than one person, then 2nd-hand sales are almost impossible? HWBot can't enforce it most of the time anyway.... most cards behave in groups of similar MHz. Bad, average, good, excellent? I can only see a problem if the second set of results were done without the component owners permission?

Thanks for posting that 9800GTX video as well

Video shows everything I would wanna see in a proof video. Ambient, GPU temp, voltage, benchmark run and MHz. Thanks for posting it up Poparamiro.... what voltages were you using for the G80 and G92 runs?

To add to that and keep the same layout: 9800GTX+ @ 972MHz Unigine DX9 Cascade average = 995MHz 8800GTS 512MB @ 950MHz 3DMark06 Water average = 855MHz. 8800 Ultra @ 810MHz Unigine DX9 Cascade average = 847MHz I have good cards and NONE of them get close to those MHz. It would be pretty sweet to see a video of a bench run at those MHz.

I am very aware that the bug has given HWB some priority work to do. Of course that comes first. http://hwbot.org/submission/2253452_stelaras_unigine_heaven___basic_preset_%28dx9%29_geforce_9800_gt_1887.36_dx9_marks If the .hwbot file contains as much info as people say, it should hopefully provide more info on how valid his results are

Didn't notice that. Err0r is not part of '501, so no crime.

The 7600GT, GTX260 and the 4770 look the same to me. 7600GT: The B1 stamp is identical- squint. Also- the soldering for where the PCI-E plug *would* be is not clean....and identical. GTX260: RAM heatsink is not covering the right edge of the memory 4770: The hand-made ink marks on the capacitors are the same. The others...... can not say one way or another.

Sure, send me one Minimum 5.6GHz please. Sometimes I work best when under pressure

Man these results are hard to follow. So.... it's an X4 960T, unlocked to only 5 cores, that show's up in CPU-Z as an X6 1600T.....which doesn't exist? Mindjob!

Friendly neighbourhood exchange rate: £500/ $780/ 595 euros

You benched too much when you're sat alone on Valentines Day evening with a hot air station, but you are happy with that decision

Got another disposable CPU This might be ok! Is/was a 3200+ Manila EDIT: I see signs of one missing component Chip works! First ever successful IHS removal!

Thanks I'd be disappointed if you WEREN'T back some day..... I used to hate Vantage so much. I think I just about got it figured out I usually have more problems with the CPU test. Wish I could run that first....

Problem created, problem had it's effect. Problem spotted, diagnosed, fix tested, fix put in place. Community was informed when there was concrete info to pass on. Other than the mistake happening in the first place, this is perfect.