Who Says A-Data DDR500 Can't Do Ratios?

[TechTones]

OK

I read numnerous reports saying that the A-Data PC4000 won't do rations like 3:2 or 5"4 only 1:1.

Well in my case that's not true. I'm running a P4P800 using the 1010 final bios.

What I did find that the ratios Won't work if the cas is left at 3.
You need to change it to cas 2.5 for it to work.

Now I did run into some problems trying a fsb of 294 and 5:4
It wouldn't boot up. I seem to top out at like 230 on the ram using 5:4

I'm running at 286 5:4 right now but my 2.6C does a 294 fsb with no problem and it has been tested with HyperX.

In fact I'm not only running it at cas 2.5, but Precharge is 3 also :)

BTW, the rammo does 275 1:1 NP at 2.85v no Vdimm mod yet. I know this stuff can do much more with more juice.

It's funny, but my CPU needs less Vcore to run Prime stable with this ram.

[eva2000]

yeah my 256MB ADATA PC4000 handle async oc's just fine as well on P4C800-E

[TechTones]

You know

I like this ram. It can be run at decent timings on an Intel board using ratios or at a very high FSB at 1:1

So it's fun to just run the system in different combos and benchmark the results.

I even ran it in a Abit Nforce2 board. OK I couldn't use cas2 but I did
manage some pretty good timings. It can be used for AMD it you don't need extreme timings. Then, when you upgrade, just use it for Intel.

I bet with 3.15v, it can do 290-295 at 1:1 which is my goal.

All I want is 294 1:1 ;)

[Creative]

Originally posted by TechTones
I like this ram. It can be run at decent timings on an Intel board using rations or at a very high FSB at 1:1

I dont mean to be pedantic or rude, but in my book decent is 2-6-3-3 as a MAXIMUM.....anything higher is not decent....2-6-3-2 at 275...now that would be decent :D
hehehehehehe

I have some of this adata sitting on the shel for over a week now cos been withouth decent cpu but should be able to test it this weekend :)

timings still matter with current day Intel rigs so offcourse this adata I expect to run at §§§§ timings but, I just want to see 275 1:1 with my owns eyes on my own rig :banana:
hehehehe

I hope you reach 300fsb TechTones :)

[TechTones]

Originally posted by Creative
I dont mean to be pedantic or rude, but in my book decent is 2-6-3-3 as a MAXIMUM.....anything higher is not decent....2-6-3-2 at 275...now that would be decent :D
hehehehehehe

I have some of this adata sitting on the shel for over a week now cos been withouth decent cpu but should be able to test it this weekend :)

timings still matter with current day Intel rigs so offcourse this adata I expect to run at §§§§ timings but, I just want to see 275 1:1 with my owns eyes on my own rig :banana:
hehehehe

I hope you reach 300fsb TechTones :)

Thank you!

DDR timings don't have a extreme inpact on mem scores with Intel though except TRCD. OK so if you do tighter timings and get 100 more points in benchmarks, it doesn't really equate into real world performance difference that is noticeable.

But I know what you mean :D

[Kalway]

As far as I've seen the CAS latency generally has the most impact on games and 3d applications. It's the Ras to Cas IIRC that usually has the least effect. Generally you want that to be 5, 7, or 11, depending on the system and RAM.

Of course that's been my experience with timings up in the lab using various chipsets. Just got an 865 chipset to try so we'll see with that next.

[TechTones]

Not on the 865/875's. Cas has very little effect.

TRCD has the most effect on mem scores and games.

[macci]

Yup Ras-to-Cas is easily the most important settings. CAS is basically meaningless..2 or 3 doesn't matter really.

Also its the Sandra benches that might make you believe its actually faster to run 1:1 with crap timings. But any real bandwidht hungry app (quake3, mark2001, aquamark etc) will tell you that 2-2-2-5 is still the way to go - even if you need to use mem divider.

[TechTones]

Yes

I observed that too Macci, but, I didn't lose that much performance by doing 1:1 probably since my 2.6c is at 3800, 3725 right now.

But I'll tell, this ram is much fun to play with. Just having ram that can do DDR500 at 2.65v is very impressive. For me it was worth it.

My P4P800 seems to love these Hynix chips. The system is fuid as hell and I can run my CPU at lower vcore's. Doesn't make sense I know :rolleyes:

[TodB]

Originally posted by eva2000
yeah my 256MB ADATA PC4000 handle async oc's just fine as well on P4C800-E

Well that's not 100% correct... see with the A-DATA DDR500 I couldn't run more than 285FSB async on my P4C800-E... no matter if it's 5:4 or 3:2... the same is with the PC3700/4000 Kingston HyperX... and all those are cas 3 default ( By SPD ).... setting cas = 2.5 doesn't help a bit....

Now when I put some cas2.5 memory, for example PC2700/PC3200 TwinMOS M.tec/Samsung and all get's put to 310-315FSB... @3:2 of course....

I mean the board "senses" somehow that the memory is cas3 by SPD and cuts me big time...

That's just what I got here people ;)

[eva2000]

Originally posted by TodB
Well that's not 100% correct... see with the A-DATA DDR500 I couldn't run more than 285FSB async on my P4C800-E... no matter if it's 5:4 or 3:2... the same is with the PC3700/4000 Kingston HyperX... and all those are cas 3 default ( By SPD ).... setting cas = 2.5 doesn't help a bit....

Now when I put some cas2.5 memory, for example PC2700/PC3200 TwinMOS M.tec/Samsung and all get's put to 310-315FSB... @3:2 of course....

I mean the board "senses" somehow that the memory is cas3 by SPD and cuts me big time...

That's just what I got here people ;) ah i see.. i think i just confirmed this with my new Q322 2.4C @ 287fsb 5:4 with ADATA PC4000 i got bios checksum heh - probably try with some of my other memory :)

[STEvil]

Wonder if there is a way to reprogram the SPD...

[TodB]

Originally posted by eva2000
ah i see.. i think i just confirmed this with my new Q322 2.4C @ 287fsb 5:4 with ADATA PC4000 i got bios checksum heh - probably try with some of my other memory :)

Yeah.... I'm trying to find some HyperX PC3500 ( cas2 ) or some good BH-5 memory.... really wanna run 5:4 with that 310FSB.... although I don't think that the BH-5 will like running 250MHz with 2.85V MAX....

[eva2000]

Originally posted by TodB
Yeah.... I'm trying to find some HyperX PC3500 ( cas2 ) or some good BH-5 memory.... really wanna run 5:4 with that 310FSB.... although I don't think that the BH-5 will like running 250MHz with 2.85V MAX.... yeah seems to doing okay with my 2x 256MB OCZ PC3700 Gold 285fsb 5:4

going to try my other PC3500 memory as well

[TechTones]

I'm at 286 now but I do seem to crap out after this. Maybe 288 at
5:4

Setting to cas 2.5 makes ALL the difference in the world for me.

To go higher try setting the so called "MAM" setting on the P4P800 to auto or standard.

Also lower the Vdimm too. In some cases this helps since the ram is not even doing DDR500 when using 5:4

[Kalway]

Originally posted by TodB
Yeah.... I'm trying to find some HyperX PC3500 ( cas2 ) or some good BH-5 memory.... really wanna run 5:4 with that 310FSB.... although I don't think that the BH-5 will like running 250MHz with 2.85V MAX....

Ewww... Yeah, 2.85v is definately a handicap. All those evil tier 1 mobo makers never give you enough volts to do some serious damage. Least MSI is starting to take their head out of their ass. HyperX loves volts and will take any amount you throw at it. My PC3000 HyperX eats em for lunch and obviously Macci's does too (Mr. 3.6v through 5ns DDR :P).

If I didn't hate the 875/865 chipsets so much I'd probably tinker with them more but 90% of them can't even pass our emissions tests, won't fit in our chassis right, or have the pci slot in the wront spot for our riser cards...

I think one of the engineers said they found a pretty sweet spot around 245 2.5-3-3-7. Took a bunch of volts to get it stable. Hardware volt modded, of course. Was on an Intel board so that was his only choice hehe.

[althes]

Just set it to 2.5438 and its working, running some test now.;)

[Creative]

Well after 5 mins testing my adata liked 275 but nothing much more currently :) I tried all the way up to 3.2 and needed 3.2 to get 275 stable. This is 1:1 mode to :)

2.5-8-4-3 was the timings I used but gave mega memtest errors...but primed a little though. I then used the spd timings of 3-8-4-4 and memtest locked up in 5 seconds....wierd

So yer, not much testing but they do work in 1:1 mode and 5:4 mode....didnt try 3:2 :)

[eva2000]

IIRC someone bigtoe or chainy said Max3 doesn't like cas 3

[althes]

currently at 275, 5:4 220 for adata. not bad at all.

[TechTones]

I top out at 287 5:4 with Mam and Turbo enabled.

If I go 288, I have to disable Mam

[o27]

I´ve managed to get 298 1:1 out of my A-data pc4000! for more specs take a look here; http://www.xtremesystems.org/forums/showthread.php?s=&threadid=18850

[NyCUndaGrounD]

Im really hating ADATA now. i have 2 sticks running 260fsb stable now. I was able to get into windows at 277fsb. I tried running it at 5:4 divider with 2.5-4-3-8 timings but is a no go. I dont think this ram can run anything besides 1:1 when running at 250mhz or more. Im giving it about 3.28 volts.

Guys is it normal when the u give it 3.2 volts and after like 10 mintues, system reboots. Are any of u able to run stable with no reboots at 3.17-3.2 volts?

[o27]

I can boot into win and run SuperPi at 300fsb 5:4 so it should be prettey stable - nb. it gets around 3.2v

[STEvil]

I've been running 3.23v for about a week now to burn in my ram good, haven't rebooted yet.