Manufactures are more than welcome to show there stuff here so please be more courteous to them when they are showing us new product capabilities.
I will be back on the RS480 soon
What is the max LDT speed seen so far? 250x5 is cake on mine.
Printable View
Manufactures are more than welcome to show there stuff here so please be more courteous to them when they are showing us new product capabilities.
I will be back on the RS480 soon
What is the max LDT speed seen so far? 250x5 is cake on mine.
I totally agree with what Fugger said. I for one am extremely happy with any info we can get on the nF4 and now Bullhead chipsets/mobos. There is no reason to criticize or bash reports about a very early beta board undergoing testing. Props to everyone who puts in the time to test and then post here. Keep up the good work. :toast:Quote:
Originally Posted by FUGGER
Oh, and as far as the LDT multis go, am I to interpret this to mean that the bullhead will be able to run at 1250+ without auto throttling back to 1000Mhz? Is this the first chipset then that can break and sustain speeds above 1000 without problems?
yup, 1250 is pretty easy ;)Quote:
Oh, and as far as the LDT multis go, am I to interpret this to mean that the bullhead will be able to run at 1250+ without auto throttling back to 1000Mhz? Is this the first chipset then that can break and sustain speeds above 1000 without problems?
Quote:
Originally Posted by bigtoe
all I've got to say is: woot! :banana: :toast:
Funny is when you read bigtoe's post with a UK accent in your head.
lol
Hopefully not to off topic here or left side, but puzzled on a couple of things.Quote:
Originally Posted by trakslacker
1 : LDT @ x5 (htt@250) logically = 1250, fine. How does a user measure the benefit of this gain? Hows does a user know for sure that hypertransport channel is running @1250?
2 : Not throttling back fine; how do we know? ( Allow me to sidetrack a little but on another recent motherboard to be release soon, x5 is also possible at 250 HTT and assumption was that it was throttling down, although bios showed it @ x5 same as here. Is it fair to assume that on the other one it does not throttle back? Tony I am sure you know which mobo I am talking about - that issue was brought up and you mentionned that it was throttling back.), hence again how do we measure that ? Are we sure it was throttling down?
3 : With Hypertransport running at 1250, but the Processor clocked at 1000 H/T do we by any chance have a false sense of increase, ie: an increased highway speed which is not utilized since the processor cannot run 1250? (or can it run 1250, 1500, etc in half duplex mode obviously or if u prefer 2500 and 3000 HT in full duplex?)
Picture this: A car with an engine tuned to run at max 100 miles an hour, it runs 100mph on a 100 miles an hour highway, when it reaches the 120 miles an hour highway it still runs at 100 miles an hour - no speed gain engine is tuned for 100 ???
Please note that these are not statements, they are in fact questions that i have and would appreciate some clarifications.
and ati finally wins the battle over nvidia in every thing.
:)
im an absolute ati fan right now.
p.d. could you please put somo screenshot comparing the dfi mobo and this bulllhead, please, is the dfi board really that far from this one in performance???
dam man sounds like physics class
blindbat, excellent questions. I'm extremely interested to see some clarifications to this.
I assume there will be no gain, there was a thread on the difference between high and low speeds done by Daegadillus IIRC. Its in the best of AMD section I think..Quote:
Originally Posted by Blindbat
Quite possible the bios could be hardlimited to say 1100 and throttles back, but i'm no bios expert and i'm sure if anyones gonna catch that its Bigtoe or one of us.. we're enthusiasts and duping us isnt the way to win around here ;)Quote:
2 : Not throttling back fine; how do we know? ( Allow me to sidetrack a little but on another recent motherboard to be release soon, x5 is also possible at 250 HTT and assumption was that it was throttling down, although bios showed it @ x5 same as here. Is it fair to assume that on the other one it does not throttle back? Tony I am sure you know which mobo I am talking about - that issue was brought up and you mentionned that it was throttling back.), hence again how do we measure that ? Are we sure it was throttling down?
They can not be run asynchronously to the best of my knowledge.Quote:
3 : With Hypertransport running at 1250, but the Processor clocked at 1000 H/T do we by any chance have a false sense of increase, ie: an increased highway speed which is not utilized since the processor cannot run 1250? (or can it run 1250, 1500, etc in half duplex mode obviously or if u prefer 2500 and 3000 HT in full duplex?)
As far as I know when its set to 5x250 the whole system runs at 1250 and not asynchronously, so the highway speed vs. max velocity of a vehicle doesnt quite fit. I do see what you are talking about, though. You might do better explaining it as latency as the distance from processor to where "HT" leads to is a point to point relationship and we are not bandwidth limited but more latency limited it seems by Daegadillus' thread.Quote:
Picture this: A car with an engine tuned to run at max 100 miles an hour, it runs 100mph on a 100 miles an hour highway, when it reaches the 120 miles an hour highway it still runs at 100 miles an hour - no speed gain engine is tuned for 100 ???
Please note that these are not statements, they are in fact questions that i have and would appreciate some clarifications.
Of course I may be off base ;)
The DFI NF4 production MB's will have a bank interleaving option in bios (so does the new NF3Quote:
Originally Posted by cantankerous
250GB Beta BTW). DFI delayed ship of the first production lots for a bios change and a small
hardware change as well appearently.
Viper
There are no tricks here. No overvoltaging, no throttling back. Why throttle back when you don't have to? ;)
I think most people have lowered expectations on HT overclocking for some reason. This level of overclocking should not come as a surprise.
Quote:
Originally Posted by trakslacker
Answer 1: Have you tried running SiSandra and see what it reports on the motherboard that is supposedly throttling back? It is usually very good at reporting the correct speeds, even on the HT link. It will probably report the correct (actual) link speed but I am not 100% sure.
Answer 2: How do you know for sure...only one way, put a scope to it. For the case of the Radeon Express 200, you can take my word for it. ;)
Answer 3: Both the CPU and NB side need to run at the same speed on the interface level.
Quote:
Originally Posted by Blindbat
http://www.xtremesystems.org/forums/...id=23263&stc=1 Got playing again with VX. Now this is just ram overclocking and i seem to be squeezing more since i played with the bios a little to make the board more friendly at boot ;)
Fugger, if you read this and want this bios pop me an email.
255x10 and pushing upto 4k unbuffered...Gotta love low latency ;)
Thank you for your time to respond, as noted in my earlier post, please understand that I am not trying to dilute the efforts of ATI or attempt at bashing it. Personaly, if ATI'S chipset was ready and delivered to the market place with performance equivalent or better then nf4.Quote:
Originally Posted by Grayskull
In a blink of an eye, I would dump nvidia, I already did for the video portion and anxious to do the same for mobo base, but that is a personal choice.
As such, I am ready to accept ''if u say so '' kind of comments, no issue there. However, I am obviously very thick and not understanding something. S'cuse me for asking again, for I fail to see the following. In the quote above, I do agree that processor and NB must run at same speed on the HT, which was my fail attempt at posting my orginial question I guess.
Question remains: Since the processor is running at 1000 what is the use of having LDT x5 @ 250 HTT?
With memory, I know we can adjust it to respond to a faster HTT, hence adjusting it's timing for example.
But for a processor there is no such tweaking possible to my knowledge, so how does one make a 1000H/T designed Processor , suddenly run @1250 or 1500 error free?
Quote:
Originally Posted by bigtoe
Dayum! :toast:
VX is looking better and better to me everyday. Bigtoe, any way you can check bandwidth for comparison with some OCZ 4400 TCCD stuff?
And thanks for the replies Grayskull, keep up teh good work everyone!
*EDIT* What volts are you running that VX at bigtoe?
Well this was the absolute max i can squeeze from my VX here on the Bullhead board.
http://www.xtremesystems.org/forums/...id=23271&stc=1
This was at 3.3Vdimm.
I hear VX speeds on the latest batch may be getting better also ;) i hope to test some soon here.
I will run my TCCD here in a while and see what i can re-tweak out of it. For now low latency has me pretty excited to it will take something special for me to start loving 300+HTT with slower timings.
When running 250X5 (1.25GHz), both the processor and NB are running at that frequency on the interface. The benefits are obvious, increase in bandwidth and if no "tricks" are played inside, a reduction in data latency coming across the link.
Quote:
Originally Posted by Blindbat
just a quick comment on the above ^
just to make sure I understand this correctly, the processor will have no problem running at say, 1250 HTT correct? Basically, just because previous mobos have maxed out at 1k HTT doesn't mean that the processor is also bound to that number.
Am I getting this right?
Right, the cpu has margin to run over 1 GHz.
Quote:
Originally Posted by trakslacker
well, awesome.
*EDIT*
I agree with the post below. I am looking to build a new gaming rig in the near future and as of now the DFI Lanparty Ultra-D is my top choice for a mobo. I don't want the SLI because I can't afford it/don't like nVidia Vid. Cards. However, as was said below, I would be willing to hold off 3 months or so if I knew I could get an ATi-powered mobo that at least equals the DFI solution as far as overclocking is concerned. Besides, I can save up more money in the extra time to buy better stuff with. So any timeframe would be greatly appreciated and could potentially earn you guys a sale from me, at least.
OT: Grayskull, I know I'm going to get a "no commment" on this, but do you have to know a timeframe for the release of the R520 cards? I've heard rumors of May-June(which would be fine), but those are just pure guesses. I'm just hoping it won't be late '05 or anything. Thanks for any info or hints you can provide. :toast:
Grayskull, is there any chance we could get a weeks / months general timeframe (as in we're looking at weeks before release or a couple months)? A lot of people are looking at purchasing DFI's NF4 solution, myself included, but if this MB is going to be out by the end of February (for instance), I'd consider waiting for it.
This information could potentially pull many more customers into the ATI camp because many of the people getting impatient (again, myself included) may consider waiting depending on the timeframe. I would guess many people are assuming months at this point since we've heard nothing to dispute it.
My best guess on ATi mobos would be definitly the first half of this year.
My question is will DFI release a board that has all the voltage options like thier SLI board does. If this happens then this might be the most kick ass board ever made.
Now I got annoyed to hell trying to get my msi k8n neo2 to run properly so i decided to go the shuttle route. SN95G5 for me. But i have seen images of a shuttle mobo on this new ATi chipset.
I don't know weather or not I will like the shuttle yet, but I hope you can get insane OC's out of that little bos like you would a full size motherboard.
The only thing i am waiting for is to hear more news on the south bridge, because all i hear now is that it is lacking any features that the current nvidia solution provides.
NOOOO. This always happens. As soon as I buy something the delivery truck is already passing me on the road with something better for cheaper than I just bought. =(Quote:
Originally Posted by bigtoe
I wonder what speed these next chips will be able to do? Any chance for an 'upgrade'?
;)