Antinomy

Don't think so. Max CPU frequency is where it all began. Max bus speed tells us a bit about the speed of the system and max memory speed tells nothing. Can you tell what is better - DDR2-800 3-2-2- or DDR2-1000 6-6-6-? The last has higher frequency it should be rewarded higher, right? I just think that removing CPU-Z was incorrect. As for other speed things - they don't tell us much about the resultive speed of the system so I agree with you.

Don't think so. Max CPU frequency is where it all began. Max bus speed tells us a bit about the speed of the system and max memory speed tells nothing. Can you tell what is better - DDR2-800 3-2-2- or DDR2-1000 6-6-6-? The last has higher frequency it should be rewarded higher, right? I just think that removing CPU-Z was incorrect. As for other speed things - they don't tell us much about the resultive speed of the system so I agree with you.

CPU-Z doesn't get boints? :eek: what the heck?

First thought - navigation became really nasty. The top border is too wide meaning you can't see any useful info without scrolling. It would be better if everything vital would be right when the screen is opened and only additional info would require scrolling the page down. Can't see team members list on the members tab in team profile - is that a bug or a feature? Can't see rankings within country, for example: http://uat.hwbot.org/country/canada Not all results are indexed, is it true? Looks like some profiles aren't in the rankings. And typos of all kinds like: but

Not only product pages for families but families as a search criteria, do you get the idea? For this one: http://hwbot.org/init.results.search.do And number of CPUs too. Say, I want to look through all dual Pentium 3 results made in PCMark. Or dual socket A and wPrime.

Massman, will it be possible to make a search not only by certain CPU/GPU model but by family or core name? Like - all E6xxx Pentiums or E4xxx Core2 and so on. It would be very handy. And person/team points update history - so you can take a look not only how much places but also how much points and from what submission did a team gain for a day/week/month. And who leaved/joined the team - meaning all team point factors should be included in the list. And not only your own team but any, like the team rankings show the place updates now. I hope implementing these two small thing won't delay the rev. for a year

Gigabyte "air MHz technology" in action

There's no problem to set Tras=1 but no advantage either Nice result!

Yeah, my girlfriend says I'm looking like an elf and doesn't allow to cut my hair so they are pretty close to the original They guy above doesn't have long hair at all.

Yep. I didn't say they should be in one month. And not even two in a row Idea was about splitting WS/Server and adding PCM05 to WS. And since we started, I remember a twin-head WS achievement suggestion. Think of it again - the current about 10 submissions is too simple

Agree - the competition doesn't give you any improvement in boints and the sharing rule was exactly about that. And this situation came up only because a GPU benchmark is ranked by CPU. It's like we've made a PCMark05 run with different videocards as stages but without dividing by CPU. So it's up to the crew to think about such a case.

Massy, you didn't get the point. For example, there can be one competition about workstation having PCMark05, wPrime 1024M. This will give 5-way for s7, s370 and s462. Another one, server could handle wPrime 1024 and/or PCM Vantage.

Each stage - different competition. One for servers (with a number of stages) and one for workstations. Otherwise some interesting platforms won't take part. 5-way stage workstation and 5-way server is more than enough. For the workstation PCMark will be interesting - assuming the highest being Socket A - there will be only two cores in the system. More than good fit for PCMark. For example, a WS competition with wPrime and PCMark (long Prime let's say). When you go to server segment - yes, PCMark05 is pretty useless. And the WS platform list is rather short (s8 is a rare one to be popular enough for the competition).

Massman, it might be a two-part competition - the ones about 771, 1366, 940, 604, 1207 are server ones. While Socket 7 (yeah, the first wide-spread mass SMP platform!), Slot 1, Socket 370, 462 are the Workstation platforms. I prefer the workstation ones but many guys like the 940, I know. And maybe a PCMark competition for the duallies?

Not gonna happen, K404. Rules aren't equal for everyone

Strike! :D You're damn lucky, Christian.

When I was against adding 3Dmark 99 and 2000 because they were loopholes allowing to fiddle with the score, nobody listened. Now you've re-discovered these bugs and started debates.

"those were the days of our lives"

Yes, Abit Softmenu was the first. Then was chaintech and pcchips. I've got a PCchips board based on sis 5597/5591 socket 7 which has voltage, multi and might be (not absolutely sure here) FSB in BIOS. Intel implemented this in their 975XBX Bad axe board say... ten years later? LOL

It's one of those dumb journalists that write what they want and claim it "the only truth" Don't believe to journalists and politicians, never. Even datasheets can be wrong while the first guys make their money on lying Why do they have the same performance? One is SATA, the other - PCI-E. Won't the viking be SATA bandwith limited? It's just a regular SSD with uncommon power distribution, no magic at all.

No, you're wrong. Like I said - I've heard of such a useless device. SteveRo has shown you a link. If you don't believe - google for IC marking: http://www.samsung.com/global/business/semiconductor/productInfo.do?fmly_id=672&partnum=K9HCG08U1M And it doesn't have the DRAM chip package either.

Yes. But for what is worth? It's an SATA SSD disk only powered from RAM slots (DDR3?). I've heard about them a couple of years ago but didn't ever see them on sale.

Christian Ney, not "labeled". Cyrix made a technology and IBM produced their own chips based on Cyrix license. So they manufactured the chips but didn't develop them. Bobnova, the 6x86 series aren't 486 clones - it's an independent technology developed by Cyrix. It was AMD who produced clones of 286-486. Only K5 was the first CPU that they developed from zero point.

IBM produced Cyrix clones. Both the one in the center and on the edge are PR200 - making them 6x86.

Nice bunch!