havli

Yes, same here - perhaps it was just a paperlaunch. Mine 5900 has Hynix 2.8ns. Running stable at 413 MHz (825 DDR), no artifacts. Most likely they can go faster but I haven't tried yet.

PCX 5900 isn't really that rare. PCX 5950 on the other hand... Mr.Scott: Thank you. Collecting hardware and overclocking is my hobby. Over the years I've managed to get many interesting things. Unfortunatelly I don't have much time to bench it lately. Most of the stuff is on my website http://hw-museum.cz/ One day I hope most of it will be propertly benchmarked and submitted here on hwbot.

Yeah, missed the background link, sorry. Now it should be ok.

Please add this board: Abit AX8 chipset is VIA K8T890 + VT8237

All these hall of fame threads are perfect, really good motivation to rebench the retro stuff and try to get the best score possible. Btw - socket 940 thread would be nice as well (since we have the socket 603/604), there are plenty of Opteron submissions. edit: one more suggestion - what is your opinion on the Cinebench R11.5? It is quite popular, reliable, multithread/multi CPU capable and all SSE2-equiped CPUs can run it. Maybe it would be good to have it in appropriate sections (socket 423, 478, 479, 603/604, 754, 939 ,940, mobile).

I hope there will be no restrictions for "one-man-team". Because I'm the only one benchmarking oldschool stuff in my team.

Yeah, GPU and CPU should be of a similar age. Couple of my thoughts... Classic 3D: 1. 3DMark 99 + Voodoo2/Banshee/Riva TNT/Savage 3D/etc. + Pentium 2/K6-2 2. 3DMark 2000/2001SE + Voodoo5/GeForce 256/Radeon 7200/Rage Fury MAXX + Pentium 3 / Athlon (SDRAM) 3. 3DMark 01/Aquamark + GeForce 3/Ti/Radeon 8500 + Pentium 4 Willamette / AXP Palomino 4. 3DMark 03/Aquamark + GeForce4 Ti /FX/Radeon 9xxx + Pentium 4 Northwood / Gallatin / Prescott / Athlon64 single-core CPU limited 3D: 1. 3DMark99 + Pentium MMX / Pentium Pro / K6 / Cyrix 6x86 + any VGA 2. 3DMark99 + Celeron A / K6-2 / K6-III + Any VGA 3. 3DMark 01 + P3 / P4 Willamette / Athlon + any VGA 4. 3DMark 05 + Socket 754 2D: 1. 486 SuperPI 1M 2. Socket7 up to K6-2 (no on-die L2 cache) SuperPI 32M 3. Dual PIII / Dual AXP + Cinebench 2003 4. P4 Willamette S423 PIFast 5. Pentium 4 RDRAM + MaxxMem Read Bandwidth 6. VIA KT133 + MaxxMem Read Bandwidth 7. EDO RAM + MaxxMem Read Bandwidth 8. Netburst CPU-Z frequency - % overclock 9. Socket 5/7 Pentium P54 (SMP allowed) - Wprime 32M Some of these could be LCC, to show the maximum effectivity (in both 2D and 3D).

Multi CPU sounds interesting. Maybe some 3dfx Voodoo benchmarking would be nice as well. I could really use some competition in that area.

Very good idea. I have tons of old HW here but not much motivation to benchmark it at the moment (except the Turrican Memorial Charity Challenge). Some kind of old-school HW league or competition would be nice.

Wow - only 9 hour... thats very good score. Different OCL driver really makes that much difference?

The first Pentium 4 - introduced at Nov 20, 2000.

Can you please add the release date to the Athlon 64 FX-51 (23rd September 2003)? http://www.cpu-world.com/CPUs/K8/AMD-Athlon%2064%20FX-51%20-%20ADAFX51CEP5AK%20%28ADAFX51BOX%29.html Also I noticed some CPUs have wrong release date assigned, for example: http://hwbot.org/submission/2740777 -> P4 3.0 Prescott was introduced 2nd February 2004 (11 years and 2 month ago is wrong) http://www.cpu-world.com/CPUs/Pentium_4/Intel-Pentium%204%203.0%20GHz%20-%20RK80546PG0801M%20%28BX80546PG3000E%29.html http://hwbot.org/submission/2739954 -> A64 3400+ release date is unknown, however I'm quite sure it was launched in 2004, not 2003. http://hwbot.org/submission/2737301 -> P4 3.4 GHz Prescott was launched at 2nd February 2004, not in late 2004 (as the "10 year and 1 months ago" suggests) http://www.cpu-world.com/CPUs/Pentium_4/Intel-Pentium%204%203.4%20GHz%20-%20RK80546PG0961M%20%28BX80546PG3400E%29.html

There are many many interesting HW platforms (mostly old pre-IMC), but only few of them are actually viable for using in the current system. For example SDRAM - only Asus CUSL2/TUSL2 (i815EP) is competitive, no point in using other boards or chipsets. In my opinion it would be nice to move "Memory Clock" plus both "MaxxMem" and "MaxxMem Read Bandwidth" to the Motherboard benchmarks. Sure - the RAM clock and performance mostly depends on the chipset used, but the actuall implementation - MB layout and BIOS affects them also. As for the IMC-equipped platforms - it is a little more difficult here... both motherboard (BIOS options, VRM, PCB layout) and CPU (IMC capabilities) affects the score. To keep things simple I guess it would be best to sort these scores by the MB as well. Perhaps the current memory sorting (SDRAM, DDR, DDR2, DDR3, etc.) could be used for global poins. What do you think?

Oldest hardware with best score is good idea for sure. But how exactly is the best overall score determined? For example - single-core CPU from 2004 finish the benchmark in 1 hour. Dual-core from 2005 scores 30 minutes... which one is better?

If SSE2 is required, then the oldest supported CPU would be Pentium 4 1.4 and 1.5 GHz Willamette socket 423. Both introduced in Q3 2000. Also - Windows 7 runs on this platform just fine, if necessary. Its time to dust off my Abit TH7 RAID.

No problem. I have checked my teemmates scores and they are all fine - default UCBench settings.

I deleted all my bugged results except this one - http://hwbot.org/submission/2306844_havli_ucbench_2011_pentium_e5700_%283.0ghz%29_465.1_mpt_score The delete button is missing, can someone of the hwbot staff remove it, please? And I am not sure about this one http://hwbot.org/submission/2569757_havli_ucbench_2011_xeon_3.2ghz_%282mb_l3%29_55.9_mpt_score It is done at default settings but still seems to be bugged.

up

Please add this MB: ASRock H61MV-ITX Thx.

I like UCbench very much - it has wide support of CPU instructions and thus scales well with many CPU generations. There are not much (multithread) benchmarks working on stuff like dual socket 5 Pentium and newest Core i7 at the same time If default settings is 100% bug-free, it would be best to keep it. I personally have no problem rebenching all my stuff to to meet the new rules.

Well, if Ucbench will be removed from HWbot, we should get some other benchmark (with HW points) as a replacement. It shoud be capable of running on Pentium class CPU (without MMX / SSE).

I just noticed this thread... and I am not sure I understand the issue completely. 1. So the "Pwds. Checked" column must contain all the numbers strictly in ascending order - i.e. the lowest number is first in the list and the highest one is last. No high numbers "in the middle". 2. The thread count could be like this? dual core CPU -> 2, 4, 8, 56, 60, 64 quad core CPU -> 4, 8, 56, 64 and how about single core CPUs? Does this bug apply for them as well? For example this score is done on single core http://hwbot.org/submission/2367371_havli_ucbench_2011_duron_1.0ghz_24.3_mpt_score Is it considered as bugged because of the Pwds. Checked number?

Antinomy: Well said, thank you. ludek111: I am not really sure. I don't have cusl2 at the moment, so I can't test this mem propertly. Last year I tried ECS P4S5A/DX+ and maximum validation was somewhere around 180 MHz as well. However SiS chipset is incompatible even with Aida64. I could only estimate RAM speed by FSB clock and memory divider. This SDRAM module is Infineon 256MB, PC 133 CL2, BX compatible (16 chip). I assume its not that good compared to modern high-density modules. Maybe BGA chips are the best.

This "wrong speed" issue only applies to the KT133 chipset (and perhaps few other VIA sdram-based chipsets using other than 1:1 mem:fsb ratio). Intel 440BX, 815 and 845 RAM speed detection works just fine in both CPU-Z and Aida64 (and therefore shows the same value). I am the only one using this platform, so I don't see a problem here. This score is 100% comparable with others. Anyway... only i815 or i845 based MB can score high enough to win this stage. I have neither of them, so I try to score as high as possible with A7V133. If mods say this score is invalid, then I will remove it and buy some CPU-Z compatible board.

Same as 1.59.