Old School is Best School Season 2 Round 1
Closed
Team Competition
Online
Old School is Best School is a team competition series for the overclockers who enjoy tuning ancient hardware. The series includes five rounds with each two stages. Each round lasts one month, so you have plenty of time to track down all hardware.
In the first round of the Old School is Best School competition series, we compete in 3DMark99 and PCMark04. In the first stage, 3DMark03, the hardware is limited to Nvidia Riva graphics cards and Slot A processors. In the second stage, PCMark04, the hardware is limited to Slot A processors. For both stages a verification screenshot and system picture are required. Good luck!
Good News ))
How many rounds planned to be in the second Season? And what about calculating winning points? separately for each round or for Whole Season? Think that for the whole season will be better, interesting and more passion. Also old HW prizes is welcomed
there will be 5 rounds. about how points are calculated id agree that it would be fun with a summary of all rounds
Last year we wanted to have an overall ranking too, but there was a code issue. Let's try to get that working fine for this season.
As for prizes: we don't have much stuff in the office, but I guess we can start a prize pool anyone can contribute too?
I'll be here! Just need some good SDRAM, CPU, and board and I'm set for this round!
A small edit here in the highlighted text.
http://oc-esports.io/#!/round/osibs_s2r1
Massman think you overclock this code issue and will find all bugs
And about prizes, it is unnecessary old-school prizes, if for example if G.Skill or smth else will be official sponsor of this Competition, modern RAM or another HW stuff bill be good for prizes.
What's the code issue?
I'll send an email to our partners to see if anyone's interested in putting up some prizes. Can't promise anything though ... Riva TNT and PCMark04 isn't particularly sexy these days
Dont get me wrong but I got really excited when I finaly found a tnt2 Ultra in a bundle with other useless GPUs yesterday! Hope the price stays realistic
Something like ... " Hey XXX company , look , your 1998 hardware came 1st , and still kick @ss " ... "Cudos , exceptional build quality"
Still a good piece of advertisement.
Non related example
Mercedes-Benz found a specific car-model here in Greece a couple of years ago.It was a taxi with a couple of million ? khm in odo ... and still running.
They took it back to their museum in Germany.
or was it an Opel
You dont really need an ultra.
There are other ways to hit the target.
So if the price gets high , just drop it.
not sure what you meen now? the ultra should have faster memory ICs alowing higher frequency atleast or maybe thats not relevant due to bad scaling?
Only checked my vanta card but assume all tnt cards takes memory voltage from the agp port as 3.3v making it a hassle to increase Vmem, Vgpu should be easy on this old cards tho.
So, background?
Oh, and Happy New Year!
Yep, backround please.
http://hwbotnews.s3.amazonaws.com/Competitions/1376-1380_OldSchoolBestSchool2016/Backgrounds/OSiBS%20S2%20Wallpaper%20R1.png
We got a background now
Guess what... found T-bird 1000MHz and Asus board. Will join soon
I'm not familiar with this platform.
Is 10x the max multi or is x10.5 possible?
I see in the interactive ocinside guide, max selectable frequency (at 100FSB) is 1050MHz, so it should be available.
Also found TNT2 Ultra (Guillemot Maxi Gamer Xentor 32) in my pile of hardware, so it seems I will be full set.
10.5x is available Ivan.
See you there.
Wow, 1Ghz T-Bird. Now I expect something crazy!
Nice competition, thanks so much making this possible.
We have a winner
Anyway, I'll try to do my best even if the chance to get good position is slim. Now... where did I put that Tbird 700..
Don't let him scare you.
Just do your best, and above all, have fun.
I've never touched slot A system, so don't be so sure
Hope kotori-san will participate too.
Plus it is all about fun
Thanks for the multiplier info Scotty.
Not sure what is possible with 1GHz T-Bird, but will try to find out.
It is going to be limited by platform I guess, but maybe 1200-1300 (at > 2V vcore) is possible with water?
Maybe. Cache divisor is the key.
KA7 is the only board that it's adjustable in the bios. Huge advantage. For all others it'll be a hardmod or software if you can find it......and it actually works.
This stage is all about the availability of hardware. GPU will have little impact as long as it's TNT2.
Run what you brung, and hope you brung enough.
Yes, however I have only M64's and one Ultra, so the latter is the obvious choice.
I have Goldfinger's device for the CPU and will try without mods first, to see how far I can push it.
And the Thunderbird has the built-in cache running at full speed, so I can't control that, right?
It is basically socket A Thunderbird on slot A board.
Correct on everything.
I have 700 T-bird also. Maybe we have a little sub contest inside a contest? :celebration:
stage 1
Theoretically, probably the guys who play with 1000MHZ cpu should have an advantage for the the stage 1,
even if they use motherboard with amd 751chiset.
The most motherboards who used chipset 751, make lets say 115 fsb on air theoretically (bios limited)
so 1150 MHZ for 1000MHZ cpu on air is possible and so on...
If your chipset is VIa kx133 ( like top motherboards for this socket, like Abit KA7 or ASUS K7V)
or any Via kx133 chipset lets say we are waiting for new world records...
chipset 751 is limited to 100 sdram and agp 1x or 2x
chipset VIa kx133 has 4x agp... (this is for the second stage advantage)
The only way if you want to have a chance for chipset 751, is gold finger device (GFD)
Northwind GFD
Ninja Micros Freespeed Pro
K7 Overclocking Card
Liberator 2 GFD
or similars like this
or you have to make it manual...
stage 2
Theoretically
PCmark 04
AGP X4 like geforce fx 5700 or 5900, geforce fx 5950 ultra or radeon 9700 pro (for VIa kx133 chipset)
pci ide or sata raid controler, and gigabyte iram or hyperdrive4
For Cache adjustments use CPUCool. Though if u use a tbird you dont need this anyway.
My TNT2 Pro just accepts 105Mhz Bus speed so I can not benefit from my KA7. Very bad...
Sure, why not.
My Tbird is pretty bad overclocker though http://hwbot.org/submission/2256175_havli_superpi___1m_athlon_700mhz_%28slot_t_bird%29_2min_35sec_964ms And I think it won't be faster than Pluto 900 MHz... let's see.
anyone know how rare Thunderbirds are compared to the cpus withseparate cashe when looking at the 1ghz version? worth taking the chanse and hope to get lucky?
1Ghz Orion runs with 333Mhz Cache while Thunderbirds Cache is 1:1. Should be important in 3D.
That I know, was a question about how rare it would be to get a thunderbird instead of a Orion when buing a random slot A 1000Mhz.
AMD Athlon (Slot A) microprocessor family
http://www.cpu-world.com/CPUs/K7/TYPE-Athlon%20%28Slot%20A%29.html
http://www.cpu-world.com/CPUs/K7/AMD-Athlon%201000%20-%20AMD-A1000MMR24B%20A.html
http://www.cpu-world.com/CPUs/K7/AMD-Athlon%201000%20-%20AMD-K7100MNR53B%20A.html
i hope that helps
Also have the innovatek oc card, but it does not work on Asus K7V rev 1.03 and older according to the description. Guess what... my board is K7V 1.01. So it does not POST.
The other problem is anything higher than 101 FSB in bios does not post.
110MHz with the jumpers works, but that's the maximum one can set.
How about 110 with jumpers and then more in win?
Consider yourself lucky enough to own this 1.0Ghz Tbird cpu and forget all the rest.
First loss of this comp.
My 700 T-bird no longer posts.
That sucks, any interesting mods done or just died randomly from normal OC?
Depends on what you call normal.
I was leaning on it pretty hard (over 2v) on the chiller.
Pretty much normal for me.
It's ok. The road to high clockspeed is littered with dead hardware.
Yes, I was extremely lucky to find that CPU and really cheap (classified as useless stuff).
I got a whole PC for like 25 euro - this T-Bird, the board, some generic sdram, quantum hdd, psu, case, cd-rom, riva tnt2 32MB.
Have no windows installed yet, but will try. Hopefully will work, otherwise it is just plain boring - set 10x110 and that's it.
Isn't 2V and above pretty normal and "safe"?
It is 18nm CPU and although not a socket it should be similar to s.A.
Thats totaly a normal OC id say, 2V doesnt sound extreme at all for a thunderbird chip (atleast not for a socket A one..), hope you got some good scores saved before it gave up atleast
@I\.nfraR\.ed, thats an amazing deal, those CPUs seems really hard to come by as they only shipped in OEM computers... I was contacted by a guy that wanted to sell one on a swedish forum, he wanted $200 for just the cpu...
I thought it was fairly normal.
Must have just been weak.
Whaaaaat?!
Actually it was over 200 euro...
Mr. Scott, have you looked at the core? These are really fragile and "chips" easily.
Haha, yea its true, no deal on that one! But must say Id like one of those after seeing Infrareds ressult in 3dmark99, those thunderbirds are amazing! Wonder if he can reach 10k even?!
Yup, core looks fine.
I'm totally familiar with all of the old school hardware. Been doing this since the mid 90's.
Found today working K7V with Pluto 750. OC from bios working fine, post up to 112mhz bus. More limited by mobo? or this shitty cpu cant do normal clocks?
Could be MB limit or the external L2 cache is holding you back.
Respect to Mr.Scott
Until right now infared is the only guy how has ATHLON 1000 Mhz cpu
For the first stage i think we have a winner here
The second stage PCMARK 2004 is tricky...
I think Ive added now all available Slot A motherboards to our db. Maybe it helps some people finding a board.
Hi, I am new here. I got this competition in my mailbox for some reason. Maybe I visited here doing some overclocking in the past. As I have some old hardware laying around here I did find a motherboard with Slot A, a processor that fits and a Riva gpucard .... can still compete with you ?
Absolutely! Get it fired up and post some scores! You have to be able to beat up on my Argon core, so many fancy Thunderbirds laying around in this competition that my CPU will make you look awesome.
Not so sure about that yet.
That sounds interesting. Bring it on!
Want to see what a true oldchool bencher can squeeze out of such system, I suck at both 3DMark99 and PCMark 2004.
I just want to make sure everyone is on the same page. We are relaxing the rule for GPUz in the screenshot? It looks like Win98 is the way to go for Stage 1 but don't want to move to that OS if the GPUz requirement is going to be enforced. Mods?
Current GPUz doesn't identify properly. According to general rules you can use Rivatuner or another program as a substitute. Should be a non-issue Tony.
http://forum.hwbot.org/showthread.php?t=148761
Its not wrong to use rivatuner but it would be useful to have aida/everest or some other app.
I'm fighting with 98 for more than one day, but can't get better score than on XP.
44.03 does not work for me, getting grey screen only when launching first test.
For some reason missed the entire Rivatuner thing. This swiss cheese memory of mine definitely kicks my butt every once in a while.
3DMark just shuts down for me when I press "Benchmark". I have no idea why, I've benchmarked 99 in Win98 before and it worked fine...
I cured this problem by "re-installing" directx. I had the same problem as you.
Tried multiple times to install a clean 98SE, but always the same - 3DMark just closes and if I try to run it again it says that I need a VGA which supports DX6.1 or newer.
Then I need a reboot.
I can't use drivers newer than 44.03 or so. Everything I get after restart is black screen with blinking cursor.
With 44.03 and some other 4x.xx drivers I get grey screen in first test. Had the same on Win2000 with 44.03.
With older drivers like 3x.xx performance is bad.
Don't really know how you guys do it
Reinstalling dx worked, thanks!
you managed to solve this? I get the exactly same problem that 44.03 just gives a gray screen in 3dmark99 and really really bad score?
windows me + dx8.1b + 44.03driver work fine, but i tested only 751 chip set
I made a fast last check before end, one result had to be blocked due to missing mem tab at 04 - good work in general
I'm just waits for surprises from Italian teams especially for stage 2 round 1...
Ill give up this round soon, my tnt2 ultra works flawless with all motherboards Ive tested it in except my k7 pro where it crashes in artifakts as soon as it get 3d load (worked the first 5-10 rounds or so in 3dmark99) at the same time that other cards Ive tested (two different gf2) works perfect in the k7 pro aswell, just makes no sense to me.
Reinstalled OS, different directX and drivers, nothing helps with the tnt2 card.
My k7pro agp does not work properly with high cpu core voltage, tnt2 ultra runs 3dmark99 limited 1.8v, 1.85v stop on the game1(race), 1.9v can't boot windows, 1.95v can't boot bios.
That kind of makes sense, the tnt2 ultra worked good until i started OC the CPU and increased the vcore to 1.85V. It does not work on stock (1.8V) even now tho.
Wierd thing is that the GF2 card works well at 1.85V all the time tho and the tnt2 works good in other motherboards.
Have you tested to change caps on it as they are like 20 years old by now?
Some caps replaced previous, but not all caps.
Any chance of extending this round-1 , for the whole weekend ???
It would be nice to have Sunday (a resting day from work) free to play.
If everybody agrees of cource.
I would be fine to that. I really want to see some great results...
I vote "yes".
I proposed this and voted in favour, we will see. Up to date, 31st of january is still valid up to further notice
You have to get up early, that's all
My EPOX board looks to have died, no idea how, was working, changed GPU, not working. Literally 20 seconds between hitting power switch on to off and won't work no matter what I've tried. Perhaps its just old, and needs a rest, will give it a try later this weekend. Good luck guys!
I had a quick discussion with @lanbonden about the end date and decided that we will keep the end date. All the competitions are up on the schedule and were announced quite a while before it started, so there should be enough time to get all the scores on time.
Good luck everyone!
Typical Epox sudden death syndrome.
Damn it, 1 minute late
Could not transfer it fast enough from benching PC.
You fail
All checked, two results removed due to insufficient verification (please do not hide benchmark settings at 99, this also includes cpu instructions used for benchmark) - good job, for me the result now is final up to possible recheck, but I think I did not overlook invalid results. Overview needs to be recalculated though, but it will not change first places
Good job
Oh no, missed the time...... Fuck
Family stuff here just made me forgot....I'm so sad now
:( Cannot believe this happened... working for over a month on this....
I am not too fond of sandbagging, I am not australian
- sad to see this, but next time it might be better to upload results a bit earlier^^
All morning was full, just slipped my mind. Was supposed to do it way earlier....I just...can't find the words...
Its not like I didnt tell him to upload yesterday evening and this morning, guess me didnt think it was a good idea to increase the competition lenght for all of sunday bite myself in the ass...
Please note that you can see all used settings in the result browser which is IMO more important than the settings in the menu as it shows the actual settings of the current run.
Yes, and this makes it double weird why people have enough space on screen and still don´t bother to show it despite the explicit screenshot at rules tab which makes it mandatory. I thought twice about this but then dicided to block these nontheless simply because rules are 100% clear, feel free to protest at head of moderation, but at 01 it is same, even if you show settings completely above the subtest window, it doesn´t match the rules. Sorry, but when I start to make exceptions and decide by "common sense" and thereby have to bend rules that are 100% clear I am not sure I will end up at making clear decisions in the future. It is annoying, I am aware of this, so I don´t make friends with this but it is the job
Too hard to work without a Goldfinger , on a Pluto cpu that you dont have a clue about it's limits.
I only had Sunday morning available (since it was not accepted to have the whole Sunday)
I soldered numerous times those tiny resistors on the athlon pcb.
But i couldnt get the result i opted for.
Regretfully good sub was miss the deadline, Marquzz-san your pcm04 is very nice! i have finished first place in stage2, but personally your win, and i was looking forward to I.nfraR.ed-san's submission.
Good scores everyone. 3DM99 was really chalenging and fun. PCM04... well this one isn't my strong point.
Looking forward to the next stage. No need for PLL mods and multiplier unlockers on s423.
Very strong PCMark04 score, indeed. Even though I have 2001 on PCMark, I lose in almost every subtest compared to kotori, so in fact my score I got just before the stage ends and could not upload so fast is worth 3rd place.
For first stage I could not solve the low score on Win98 and on Win XP I could not improve much, since the board was not stable at 112 and 115MHz no matter what I did.
The goldfinger device somehow worked and I also tried to force 10.5x multi by desoldering one resistor, but it did not boot. Same thing if set with GFD.
This CPU has no VID3 resistors, so I don't know what else to try. The CPU is definitely capable of higher frequency, but I was limited by the board.
I also got an Epox board yesterday evening, but could not run the 1GHz Thunderbird at higher than 8.5 multi and the board was so unstable even on stock...
Guess I need TurboPLL to achieve more, or maybe some better board, like the Abit.
Good job everyone, kotori-san was unreachable for me.
-------------------------
My GFD has 12.5x setting but i also can't boot up 10.5x multi.
About EP-7KXA, I.nfraR.ed-san do you have Argon or Pluto athlon? i found bug report in japnese over clocker home page, roughly matches your symptoms, problem only using with T-bird (Aluminum core, Copper core both).
Other than that, sorry japanese web site.
http://ascii.jp/elem/000/000/777/777911/index-3.html#eid777918
Translated a part
It is a story that was published at the time EPoX support site has the following of such circumstances. The Athlon there is a large two signal mode. One is a method called Open Drain, the other is a Push-Pull method.
Neither is a method of well-used, but simply put Open Drain is a signal of how to cover the High and Low is one of the FET, a method of allocating another FET in the generation of Push-Pull The High and Low. It is preferred that more of the Push-pull is when considering the speed of the signal, founder of Athlon Slot A (K7 / K75.K76 unknown) had been constructed in the Open-Drain system.
On the other hand, in the case of Thunderbird, but all Socket A is a Push-pull method, Thunderbird for Slot A was possible to Open-Drain output. However, to do this requires a special setting for instructing a "case and the Open-Drain output" to the Thunderbird, which was not only implemented in the AMD-751.
When mounting the Thunderbird to KX133, against the intention of CPU to transmit and receive signals in the Push-pull, would have meant to send a signal in the Open-Drain system, this is the CPU on the grounds driver (transistor part for transmitting and receiving) but there is a risk of damage.
Yes, I have a Argon 650MHz and it booted straight up.
Don't like the Epox board though, better mod the Asus, which is much more stable.
Seems like AMD said that the Thunderbird had no support on KX133 boards, despite the fact that those where released later. But AMD chipset had correct support for it. Only exception, was Asus K7V-T where T stands for Thunderbird. Toms Hardware released a guide how to upgrade K7V to a K7V-T here: http://www.tomshardware.de/aufpoliert-vom-asus-k7v-zum-k7v-t-uning,testberichte-209.html (german)
Its too bad some of you missed the deadline, but I would wish you would upload your results anyway. Just to show what this platform is capable of.
Another KX133 board, Aopen AK72 was perfectly stable on Thunderbird.
@Gumanoid
Since we used the same board (Gigabyte GA-7IXE) , i wanted to ask you a couple of things , in case you happen to know , or you actually faced the same problems.
1)With a hardware modded Pluto @1.85 & 1.90 volt , the board refused to boot.
2)Having a rev 1.1 board , i couldnt upgrade to bios F7 official or FAA beta.
When i flashed to those two bios versions , the board was dead , not booting.
The only operating bios was the F3 (i need to check-out the F4 beta).
3)Hardware modded cpu @1.80 volt.
When reading pll output with a software prog. on my next boot i was having 2.05 volt Vcore !!!
My K7V-RM runs Thunderbird just fine. KX133 board.
@Strunkenbold, @Kojima45
Very good info, thanks.
Will try it in the future. Want to see the limit of this CPU.
External PLL mod does not seem too hard to make, so I will try. Never done it before, but love to learn new things.
I also want to see.
@scotty: K7V-RM looks good, i'll get fi find it.
Yes, my board is rev 1.1 too, my board is no problems with flashing to F7 bios. Where did you get a FAA beta bios?
Unfortunately I have not tried to modify the vcore.
Maybe it's my board , that acts strange.
I tried multiple times to flash to a newer bios bersion and everytime i did , the board was not booting afterwards.
I found the latest beta on a webarchive page.
I can give it to you , if you want it.
Log in or register to comment.