Gigabyte shows off i-RAM successor

[Makaveli]

Show me a benchmark thats proves u get a 30% memory speed hit when going to 2t from 1t please?

And u might want to re-read my post. there is a reason I quoted what he said. I didn't make any comments about the I drive. it was strickly talking about that non stability statement.

Also the avg HD does not have a avg 70mb write speed. Maybe if your taking about WD raptors and the newer 400+ GB HD's anything smaller on 7200rpm does not push that much speed.

and stop quoting the bandwidth speeds of SATA and SATA 2 for the Idrive show me a benchmark with it pushing that speed please.

 

[zekrahminator]

I find it hilarious that we can have such heated debates over computer parts .

 

[Makaveli]

This isn't a heated debated at all, I just want proof to his claims which I think are false.

 

[Mussels]

Show me a benchmark thats proves u get a 30% memory speed hit when going to 2t from 1t please?

And u might want to re-read my post. there is a reason I quoted what he said. I didn't make any comments about the I drive. it was strickly talking about that non stability statement.

Also the avg HD does not have a avg 70mb write speed. Maybe if your taking about WD raptors and the newer 400+ GB HD's anything smaller on 7200rpm does not push that much speed.

and stop quoting the bandwidth speeds of SATA and SATA 2 for the Idrive show me a benchmark with it pushing that speed please.

Lavalys everest ultimate. I cant be arsed, but if you run an AMD64 system and swap from 1T to 2T, watch the write test. The worst i got (with bad latencies too) was a 28% decrease. That was running 4x512 dimms on an Nforce 4 ultra. 4x512 didnt work at 1T, but at 2T was a ton slower than 2x512 at 1T

My claims are not false nor 'heated' its an old topic to me.


Edit: said read test, meant write test.

 

[Mussels]

This isn't a heated debated at all, I just want proof to his claims which I think are false.

i dont like being called a liar.

Edit: just so that people know, the 2T picture blatantly shows 'CR2' which means 2T timing is in effect. DDR speed as well as other timings are visible and un-changed between tests.

29% performance drop, case closed.

Edit 2: I forgot to say, these are on 2x1GB sticks, not the 4x512 i mentioned earlier.

 

[W1zzard]

gigabyte just told me it does not use sata-ii. the problem is that the xilinx chip can not run that fast. this is because gigabyte has to engineer a memory controller like in a chipset, without previous experience (like intel, amd or ati) this is very hard to do .. 2gb modules do not work yet either but its definitely planned

 

[Mussels]

Are you sure it doesnt just use SATA 300, instead of SATA-II (aka 300MB, but without NCQ etc) ?

 

[Makaveli]

U got to be kidding, Show me Real world benchmarks with real applications not syntheic junk

Like sisoft sanda and everest, i'm not some noob, that might work on some of teh newbies but not me.

case it not close, u haven't proved anything.

Show me a game taking a 30% hit running 2t?

i'm not asking for alot.

 

[Mussels]

If that is all you have to say, i will not bother replying to you. You asked for proof of a performance hit, i have shown it. The difference will be less in 3D apps but we are not talking about that. This is NOT a discussion about 3D performance, simply about the fact that running 2T is slower than running 1T. THAT is fact.

I am not discussing this further, as you seem to not give a shit and wish to live in your own little world. Let's allow this topic to get back on track.

 

[Makaveli]

That is the reason I called u out on it u have no proof. People actually use there computers for applications and games.

I don't know about u but I don't play everest and Si soft sandra. I use real applications.

that fact that u can't show me a real world appication seeing a 30% hit just shows i'm correct.

Thanks for coming out. The only one living in a little world is u my friend.

 

[Makaveli]

If there is anyone actually interested in some real world data about the difference from 1t vs 2t on athlon64 rig.

here is a great article showing what i'm talking about.

http://www.xbitlabs.com/articles/memory/print/2gb-ram.html

Just found another good link

http://techreport.com/etc/2005q4/mem-latency/index.x?pg=1

enjoy the read, maybe you will learn something.

 

[CjStaal]

Yes, I believe the OS is on the ram disk.

Also even when you completely turn off the page file using that method Windows still makes and uses one. If you don't believe me then try it, turn off the page file then reboot and go into task manager. Under Performance it will still say Windows is using a Page File. Windows really does not like not having a page file, you can force it to completely not use a page file using tweaks, but then it gets really upset about it.

That is RAM +PageFile dude... it just says pagefile even if their isn't one... why don't any noobs get this?

 

[newtekie1]

That is RAM +PageFile dude... it just says pagefile even if their isn't one... why don't any noobs get this?

No, it isn't. My current memory usage is 571MB, my PF usage is 498. If it was Ram+PageFile the number would be equal to or bigger then 517, not smaller. Nice try though n00b.

 

[newtekie1]

U got to be kidding, Show me Real world benchmarks with real applications not syntheic junk

Like sisoft sanda and everest, i'm not some noob, that might work on some of teh newbies but not me.

case it not close, u haven't proved anything.

Show me a game taking a 30% hit running 2t?

i'm not asking for alot.

First of all he never said games take a 30% performance hit, he said memory write speeds take a 30% hit which he proved. Game do take a performance hit from this and it can be noticeable even if it isn't 30%. Not to mention that 4 memory modules usually leads to worse overclocks and the need for RAM dividers which can again drastically affect performance.

http://www.anandtech.com/memory/showdoc.aspx?i=2560&p=2

You can read there for more information.

The case is close, he proved that is original statement is correct, and I just proved that it then translates to a gaming performance hit. If you are going to call someone out on something at least make sure you know what you are calling them out on.

Also, my statement about more then 2GB of RAM making A64 rigs buggy was made because there are almost no 2GB sticks that have decent latencies and that are not ECC, both of which are performance killers. So the only real option a gamer has for more then 2GB of memory is to use 4 sticks, which is why I made my statement that over 2GB of memory and A64 rigs get buggy. If you use 2 sticks then you have to deal with the slow-down of ECC and high latencies, if you use 4 you have to deal with the slow-down of 2T and worse overclocks. Either way you go more then 2GB requires you to make other sacrifices.

 

[KennyT772]

ok for newtekie. open up the task manager (ctrl + alt + del incase u didnt know). on the very bottom of the box there will be three peices of info. for me this is
Processes :30 | CPU Usage: 0% | Commit Charge: 328/1927
the last peice of info is the amount used/amount total of space for programs to use, aka ram (1024mb) + pagefile (1000mb-1500mb). when you turn the pagefile off it will show somewhere around (current ram load)/1024-1050 as windows needs some hdd space for system temp files and things that cannot be lost if power is lost before shutdown. no matter what you do there will always be a pagefile in windows. the only thing you can do is set weather windows will use it or not.

now on the 1t - 2t question. 2t causes a hit in performance end of story. in memory only benchmarks it can be up to a 30% hit end of story. will there ever be a 30% hit in app performance or games? no end of story. you dont seem to get that when you are running a game more things impact it then just you command rate. but i garentee if u have a 2500+ running with a x1900xtx and set everything at lowest grfx settings it will make a hell of a lot more difference then if u crack the res all the way up and turn everything on max.

now on the more then 2gb makes a a64 buggy question. in reality it doesnt matter what system u have today more then 2gb will get buggy end of story. there are not good timings yet on 2gb sticks. also if u run 4x512 its still gonna be buggy due to lots of despersed memory addresses, and maby one of the mem chips is a litter slower then the rest on the stick.

so before you call anyone a n00b or newb or noob just shut the fuck up as u dont know what you are talking about. if u did you wouldnt bitch at those who know less then you.

 

[Makaveli]

wow talk about getting it twisted, I know there is a hit when going to 2t that is obvious my point was it isn't as high as he stated in real World applications. Show me w real world application taking a 30% memory write hit. And not a si soft or Everest benchmark.

Which he didn't so I still stand by my post!


And kenny please explain to me what your definition of buggy is? I do know some users with 4x1GB sticks in athlon 64 system that are as stable as my 2GB system. u care to explain why thier rigs aren't buggy.

And if you were to re read my post I never once called him a liar or a newb.

I just wanted proof of his claims.

 

[Makaveli]

also smartass, if your gonna give an example alteast make it a usable one.

You can't run a x1900xtx in a AGP motherboard first of all.

And if you happen to be lucky enough to own one these cards who the hell would be playing at the lowest possible settings anyways.

I think your the one that needs to STFU.

 

[newtekie1]

wow talk about getting it twisted, I know there is a hit when going to 2t that is obvious my point was it isn't as high as he stated in real World applications. Show me w real world application taking a 30% memory write hit. And not a si soft or Everest benchmark.

Which he didn't so I still stand by my post!


And kenny please explain to me what your definition of buggy is? I do know some users with 4x1GB sticks in athlon 64 system that are as stable as my 2GB system. u care to explain why thier rigs aren't buggy.

And if you were to re read my post I never once called him a liar or a newb.

I just wanted proof of his claims.

When exactly did he say how big of a real world hit it is? Oh wait...thats right. HE NEVER DID! He just said it did make a difference. Why are you arguing with this? Again he backed his original statement up 100%.

Buggy, in my definition, is anything undesired. It don't have to be unstable, it is unstable at 1T, and forcing me to run at 2T is something undesired. Buggy does not always mean unstable. Oh and besides that, why are we forced to run at 2T? Because it is unstable at 1T, so even your "buggy has to mean unstable" argument is fulfilled.

Oh and by saying you think his statements are false, you are calling him a liar. To bad he backed up his statement word for word.

 

[Mussels]

Makaveli: You are just making an idiot of yourself. I never once stated real-world performance loss, that was you. As has been said i backed up my statement that there was a loss, without going to the extreme lengths of doing 20 odd gaming tests for you. This thread is NOT about games. Its about the advantages of the I-ram have memory in it, compared to the main system.

Do not call that guy a noob over the XP2500/1900xtx example. IT WAS AN EXAMPLE. He matched an old CPU with the fastest vide card around, for the specific purpose of making a point. He did not suggest they were compatible in the real world. He did not tell you to buy that combination, so please just shut the hell up and stop whining.

 

[Makaveli]

wow you guys have selective reading, when did I call that other jackass a noob.

 

[Mussels]

and you have a selective brain, to find the slightest flaw in what we say so that you can distract people from the things you're doing and saying. Noob, jackass... you're calling people names constantly.

Hmm you didnt happen to have edited your post did you?
Last edited by Makaveli : 06-09-2006 at 11:07 PM.


"And if you were to re read my post I never once called him a liar or a newb.

I just wanted proof of his claims."

If you want us to beleive you, dont edit your damn posts.

 

[newtekie1]

wow you guys have selective reading, when did I call that other jackass a noob.

Well lets see, so far you have implied that I have learned my computer knowledge from the back of a cereal box because more then 2GB of memory, according to you, doesn't make A64 systems buggy...though I already pointed out that it does in fact give undesirable results which to me equals buggy and that 4 sticks of ram, which is requires to run more then 2GB of memory and still keep dual channel and every other performance enhancing feature would always run unstable...so still even your "it isn't unstable so it isn't buggy" argument is out, because it is in fact unstable with more then 2GB if I want to keep the same performance.

Then there is the whole smartass, and obviously the jackass remarks about the other poster, not to mention that implying that people are liars and then trying to put words in their mouth to try and make it look like you are right. So, while not directly saying n00b, you have resourted to childish name calling because you probably realized that you know the least out of anyone in this discussion about the subjects at hand.

 

[Mussels]

Well lets see, so far you have implied that I have learned my computer knowledge from the back of a cereal box because more then 2GB of memory, according to you, doesn't make A64 systems buggy...though I already pointed out that it does in fact give undesirable results which to me equals buggy and that 4 sticks of ram, which is requires to run more then 2GB of memory and still keep dual channel and every other performance enhancing feature would always run unstable...so still even your "it isn't unstable so it isn't buggy" argument is out, because it is in fact unstable with more then 2GB if I want to keep the same performance.

Then there is the whole smartass, and obviously the jackass remarks about the other poster, not to mention that implying that people are liars and then trying to put words in their mouth to try and make it look like you are right. So, while not directly saying n00b, you have resourted to childish name calling because you probably realized that you know the least out of anyone in this discussion about the subjects at hand.

Especially since 1T/2T isnt even what the discussion is meant to be about, but the I-ram 2...

 

[Jimmy 2004]

also smartass, if your gonna give an example alteast make it a usable one.

You can't run a x1900xtx in a AGP motherboard first of all.

And if you happen to be lucky enough to own one these cards who the hell would be playing at the lowest possible settings anyways.

I think your the one that needs to STFU.

WTF? Ok smartass, when did he say it was an AGP card?

 

[Makaveli]

jimmy he said a 2500+ and a x1900XTK . Guess what there is no 2500+ a64, therefore he was implying a Athlon XP 2500+ which is socket A, there for a AGP motherboard. jesus christ are u guys all blind.