TPMW T=trash no way!! stepping myth

I have been reading latetly that this stepping does this much overclock and that steeping does that. I think it is very hard to just go by stepping. I have a seen a few people bashing tpmw mobiles by saying they are trash. I think it is safe to assume that you can get a generaliztion of how much you might be able to overclock with a certain stepping but it is never for sure.

For instance i bought 2 mobile 2500 with the following stepping
aqyha0410tpmw. Also the 3 digit number at the end of the second line which i have read that the lower is the better is as follow 112 and 014.
The one with 014 at the end maxed out at a little under 2.4. It would post and load into window all the way up to 2.7 but i do not consider that because it would not pass prime95 till it was at least below 2.4 with about 1.8vcore.

Now for the other with it ending with 112 was a good one. It could get it to post all the way upto 2.8 but was not stable till it was under 2.55 with 1.675 vcore. One thing i did notice was the mobiles do not like to operate near 50c. If i gave it too much vcore and the temp got over 50 the screen would lock up.

So my final stable overclock was 12x210=2525 with 1.675 vcore and prime95 with 22 hours. So i think the lesson too learn is you can not always tell what a stepping is going to do till you JUST OVERCLOCK IT!!!

here yo go

I think its more due to inexperience than bad hardware. On that note... CONGRATS!

Thats the OC im shooting for

This issue has been discussed several times in the past - as far as I am concerned the most important factor is week of manufacture.
The rest of the stuff is perhaps related to what frequency rating it passed and at what voltage and maybe though unlikely elements of incoming "wafer integrity" perhaps shown on the batch card for the incoming product from Dresden.

Consider this scenario:

Testing, Marking and Packaging plant in Penang, Malaysia

AMD employee 1 (Encik Rashid) "apa khabar, Encik Juhari? "
AMD employee 2 (Encik Juhari) "khabar baik!"
Thats basically:
Hows it hanging dude?
Sweet thanks
in Malay
Anyway back to the storyline
Employee 1 "Hey Juhari - I have this great idea about marking the cpus!"
Employee 2 "Whats that Rashid"?
Employee 1 "Why dont we test the limits of the cpu and when we know how far it overclocks we give it a secret letter hidden within our marking system to show if its a great overclocker or a poor one!"
Employee 2 "Cool, great idea but why does that benefit us?"
Employee 1 "It doesnt but it would be fun wouldnt it!"
Employee 2 "I guess so, how long would it take to do?"
Employee 1 "I reckon no longer than 5 minutes per cpu!"
Employee 2 "OK thats not long so it should be no problem - how many cpu's do we test in a year?"
Employee 1 "About 7 million"
Employee 2 "But that means we need to test about 20000 a day or 1 every 5 seconds that means we have to do 60 in 5 minutes not 1!"
Employee 1 "Oh I never thought about that".......

Bottom line when the cpus are tested they are likely frequency tested at the highest possible frequency e.g. for Desktop Bartons they are tested at XP3200 ratings and at 1.65V if they pass then they can be marked as any particular Barton in the range depending on product demand. If they fail (more and more unlikely with such a mature product - hence the introduction of low power Bartons) they are tested again at XP3000 until it passes or is scrapped. They might also be tested at lower voltages - but that is the extent of the testing that is carried out that might infuence overclockability.

Hah. Good one.

I'll be quoting that one in hopes of getting a retort but I think it'll be a while, I think the best they'll do is probably, "what, you think those are just some random letters...?"

lol

If I have all the time in this world, I would like to get one for FREE if AMD wants me to test all of them

I'm starting to think that the 4 letters represents the days the CPU was made..since they've been like XPMW, TPMW etc....
Could be the silicon that was produced on a particular day being used for that week.......Just a hunch..

The P and W are always constant the other two letters change however.

jkb2198 said:

here yo go

Click to expand...

What kind of memory do you have and what voltage are you running it at? What about the timings for it? t hanks.

Still trash when compared to MPMW. 2.5Ghz is nothing realy, childs play. It didnt do 2.7Ghz, did it? My mpmw does 2.5Ghz @ 1.65v, and goes all the way up to 2.7ghz.

You can hardly call it trash!!!
By your own admission you dont use Prime 95 to determine stability whereas he does - and he stated it gets to 2.525GHz with 1.675v you need 1.9v to get to 2.7GHz - so you are not making a convincing point.

Y dont people enjoy what little they have-get instead of complaining about what the dont have-get?

No problems with mine. It is stable to 2.6.

Very good point oc detective. Earlier today i had just read one of gogeta threads and notices he tests stabilty with games!!!!
Games do stress your computer but not near as much as prime95 or folding@home. I am stable at 12x210=2525 at 1.675. i think this chip can go higher as it will post all the way up to 2.8 and load into windows with only 1.75vcore. Now if it is stable i do not know. This is also my water cooled rig and i dont like to see high temps. That is why i will probably keep it where it is right now. As my idle temps are 40c and it maxs out at 45c. Try prime95 gogeta then ge back to us with a screen shot.

Also oc detective could you tell me how to get the images i capture smaller so i can post my screens or tell me what you use to post your screen shots.So i can show you guys. Thanks just pm me

Keep clockin

Everyone has their own methods of testing stability, their own standards, etc. and I can respect those.

But you can't compare one measure of stability to another.

Just because one processor is "gaming" stable at a certain speed, doesn't mean that its better than another processor at a slower speed that's Prime95 stable. They aren't comparable by those means.

And as for Gogeta's own stability measure:

Gogeta_G3D said:

- 2v, 14x200
Made it into windows, but crashed immediatly. I think it was getting too hot, or even rail was low?

Click to expand...

You yourself stated that 2.7GHz wasn't even remotely stable by any measure, yet now you claim you've attained the same overclock at only 1.9v?

Every time one of these letter arguments is made, its always shown that they don't mean a tremendous amount, if any at all. I know, as I've tried...

to make it easy for everyone, it is a part of the date code.

from here:

I got one of the very first batches of Moblie 2400's that were at newegg. Running [email protected] and I need 2 volts to run at 2.6-2.63GHZ. But I am really not complaining B/C what is the fastest retail barton at? 2.2GHZ, so I am still way ahead of the game. Trust me this thing flys. Stop whinning if one only does 2.4 or other do 2.6-2.7. Be happy with what you have.

and

btw, this is from this page

http://www.vr-zone.com/?i=746&s=4

I am just trying to prove the point that going by stepping is really hard to tell what you are going to get out of a processor. I am not whinning in any way and i am really happy with the performance i am getting out of the two mobiles i own.

My other point is that people say my mobile overclocked this much so mine is better than yours. With most of the time not even testing it for stability. Which is hurting everyone at this forum because your results are not totally valid. People just get it to load into windows and get a screen shot saying look at what my mobile can do.

Also stability for me is not just based on one program but multiple programs. I first test with prime95 to see if it can pass for 24 hours. Than i use memtest86 and let it run for 24 hours. Then i loop 3dmark for 12 hours and see if it will crash. Only after this do i deem a overlcock stable and then share my results with the forum. It is hard to test for stability because programs such as prime95 do not test just the cpu and memtest does not just test the memory. I just fell if you can run three very demanding stability tests and pass all your overclock is very stable.


And once again my results where as follow to make it clear to everyone.

This processor will load into windows all the way upto 2.8 with a vcore 1.8 but i have not tested it for stability that high.
Right now i am at 12x210=2525 and had to bump up my vcore to 1.7 as yesterday morning after doing prime95 for 24hrs it failed 3dmark on loop after 9 and a half hours. After bumping up the vcore it passed looping 3dmark for 12 hours. After passing prime95 for 24 hours and memtest86.

Thanks for all you guys opinions as this helps us become better overclockers.

Keep clockin