DFI RDX200 CF-DR First look

Printable View

Show 100 post(s) from this thread on one page

10-19-2005, 02:37 PM
s7e9h3n

Looks as if OPB can't get his board to go over 235mhz with his memory. Also, he's unable to run spi @ any speed. Anybody got a link to an updated bios? Or Clockgen for this chipset?
10-19-2005, 02:57 PM
perkam

s7, link to opb's results ?? Those results look very discouraging.

Perkam
10-19-2005, 03:07 PM
s7e9h3n

Quote:

Originally Posted by perkam

s7, link to opb's results ?? Those results look very discouraging.

Perkam

Hehe...i just got off the phone with him-I think he's too embarrassed to post :p: j/k No screens yet, he's gonna talk to Oskar first ;)
10-19-2005, 03:37 PM
eva2000

Quote:

Originally Posted by Prometheus

Always nice to hear from you, Eva. I've missed your thoughtful questions since the old days at Asuboards/ABXZone/BleedinEdge but I still see a comment from you once in a while at AT.

Oskar made improvements to the board after the initial shipment to the US DFI office. He was still tweaking the boards in Taiwan as late as Wednesday of last week. The revisions were locked last Wednesday and my sample was shipped direct from Taiwan. Overnight from Taiwan takes 2 days to the US and UK. We received boards on Friday afternoon for me, and Friday evening for Tony.

In the meantime DFI California was modifying boards to the final Oskar specs for delivery to New Egg, who was to begin sale of the RDX200 on Monday. This is unusual as we usually have at least a week or two to test new launches like this board, but I worked all weekend and slept 3 hours Sunday night to get the major testing done and the review written and to our web editor. The board is super stable, but the memory timings and options are not at all like the DFI nF4. It took a lot of calls, Oskar Saturday night, Tony, and RGone to get things working properly to bring you decent info in such a short period.

Frankly, the AT audience is not really xtreme - I often refer those readers here to xtreme in my reviews - but I still try to bring you at least the OC basics which can provide a starting point for going much deeper as most of you do. That means I will not likely do a follow-up Part 2 on this board very soon.

Thanks Prom, guess we need more input and results from Tony and Rgone as to how particular memory run on this board and their settings for performance and stability :)

But guess there maybe more improvements already ATI lists a Pure Innovation Crossfire II coming soon http://www.xtremesystems.org/forums/...75#post1078175 SATA2/better USB performance ???
10-19-2005, 03:52 PM
althes

Does anybody else have the board so we can more thoughts from other people.
10-19-2005, 04:37 PM
Haltech

Im sitting this round out. This is ATI's first stab in this market... Im sure they will get it right on the second attempt. I just find it nuts to pay $240-$260 shipped for a board lacking important features in the southbridge. Thanks Tony for providing this info for everyone and putting up a pretty good fight to defend ATI and DFI in this manner. Its helped make my decision easier.
10-19-2005, 04:40 PM
4rory

Just thinking out loud here but, what would it take to get sli working on one of these boards?

I'f im well informed sli doesn't need anything chipset wise like the crossfire.

If this board could run sli, ... that would be very nice for alot of reasons.

Thanks tony for answering my questions.
10-19-2005, 04:52 PM
saaya

hey wes!
nice to see you post again, been a while :D
thx for taking your time to post here, and sorry jason got so upset about things :D

jason, i think your getting to emotional with this again :D
i get your point and partly agree with you, but its too early to judge this board and imo you shouldnt get so upset about this.

i removed parts of some posts wich were imo just provoking and didnt add anything to the discussion. please stay on topic, and if you want to post your opinion about it whether its possitive or negative, please do so without offending or provokig anybody else.

if you dont like something just say so and dont blame anybody for it, thx :)

now regarding what you said about usb and hdd performence wes, i disagree with you there. you said most people wont notice it and wont care, i my opinion thats not true, at least not totally.
if you have two competing products and one is just a tiny bit faster or lacks a tiny feature the other one has people will go for the better product, even if the difference is incredibly small.
thats my expirience at least.

whats confusing me is not that the board has lower usb and ide performence, what confuses me is that it was probably possible to perform the same or even outperform the competing solution without much work at all, but for some reason it didnt happen.

and i have to agree with jason that the review was really very sort and only shows some few results, id rather have called it a preview and not review.
thx a lot for working so hard on it though to bring us at least some basic infos before the board reaches newegg! :toast:

i hope you will focus on memory tweaking and crossfire performence in part2 of this review, wich is imo the 2 interesting features of this board.

tx a lot to wes and tony again for answering all those questions! :toast:
10-19-2005, 04:59 PM
saaya

Quote:

Originally Posted by 4rory

I'f im well informed sli doesn't need anything chipset wise like the crossfire.

sli needs driver support...
afaik crossfire doesnt need any support from the chipset, all you need is to be able to mirror the data sent to the videocard so both cards get the data afaik.

crossfire should work on nf4 boards and sli should work on crossfire boards afaik, its all a driver thing, so in the end its up to nvidia and ati.

but thats the problem...
since nvidia already has a load of sli boards and sli cards out in the market they should rather be interested in not allowing crossfire on their boards and not allowing sli on atis boards. that way people with an sli board will have to get sli cards, and people with sli cards will need an sli board, so more people buy nvidia...

so i think the chance for running crossfire on an sli board or sli on a crossfire board are low :/
10-19-2005, 05:03 PM
Tony

Guys you need to measure performance of sata2 on the NVraid, then measure IDE, then measure or compare against SB450 and the ULI1575...you will see ATI win some, ULI win some and NV win none.

Issue is no one is moaning about the NV raid though which kinda confuses me.
10-19-2005, 05:43 PM
Ugly n Grey

NVRAID? gotcha... first thing I turned off... gotta have lots of BIOS space for real RAID ...may I calculate that XOR bit for you Sir? Cache some data on a real DIMM ? Thank you sir. LSI Megaraid anyone?

Buffoonery aside, everyone complains about NV IDE performance, that's why there is a bunch of threads asking about uninstalling the drivers from Nvidia and going back to the default windows drivers to try and gain at least a little performance on it.

So yeah, people complain.
10-19-2005, 05:55 PM
Prometheus

Quote:

Originally Posted by saaya

whats confusing me is not that the board has lower usb and ide performence, what confuses me is that it was probably possible to perform the same or even outperform the competing solution without much work at all, but for some reason it didnt happen.

The ATI SB450 actually has BETTER IDE and SATA performance than the nVidia nForce4. Check my results at http://www.anandtech.com/cpuchipsets...spx?i=2542&p=9. The only real issues is USB, which is less than half the throughput of nVidia. Sata2 and NCQ are not there on SB450 and I totally agree they should be there to compete with nForc4.

I was merely pointing out that performance wise, SATA2 3.0Gb/s and NCQ really make no performance difference right now. SATA sustained transfer rates are still maxing out at around 100MBps (0.8Gbps), and that's only with RAID-0 and two drives... so SATA1 can easily handle current real transfer rates with it's 1.5Gb/s. NCQ is similarly not very important at present.

It's like nVidia's new Dual x16 SLI. We're all testing it but we can't find any test results that show it makes any performance difference at all over dual x8, but we're still looking.
10-19-2005, 06:52 PM
T_M

Quote:

Originally Posted by Tony

Somthing you guys don't know is when the NF4 was lauched by DFI the bios it shipped with was a CPU killer, i killed an FX53 and Rgone killed a cpu also within hrs of us first testing. We brought this to DFI's attention and numerous bios files were written to eliminate the issue and bring performance back.

Not laughing at the fact you two lost CPU's, but i find it hilarious that this can happen. Do they not test the boards themselves first, or do they just design, manufacture, and ship? :stick:
10-19-2005, 07:01 PM
Ugly n Grey

The problem with the SATA numbers game is people somehow believe that the bigger number makes a difference. Sustained write performance never enters the equation. The actual internal transfer speed to interface of the drive is the choke point and only the cache bursts at anything resembling those numbers. But 3.0 is better than 1.5 right? Must be, it's a bigger number after all.

Would you rather have a 250hp motor in a 2000lb car or a 350hp motor in a 3500 lb car if your goal is speed?.... Bigger is not always better. The numbers are decieving...

/rant mode

UnG
10-19-2005, 07:03 PM
IvanAndreevich

What's new, Grey. It's always like that in the computer industry. Numbers rule! However, the new drives (SATA2) are usually otherwise updated as well (higher platter density, quieter motors, etc). Seagate 7200.9, Hitachi 7T250..
10-19-2005, 07:03 PM
eva2000

Quote:

Originally Posted by Tony

Guys you need to measure performance of sata2 on the NVraid, then measure IDE, then measure or compare against SB450 and the ULI1575...you will see ATI win some, ULI win some and NV win none.

Issue is no one is moaning about the NV raid though which kinda confuses me.

Tony you sure ? i totally destroyed titac's posted UL1575 SATA benchmarks at http://www.xtremesystems.org/forums/...77#post1078177

and

http://www.xtremesystems.org/forums/...38#post1078238

i know it's 2x and 4x Hitachi 7K80 SATAII raid 0 but wonder how same 4 drives handle on SB450 and UL1575 ? as that's what i run nowadays 4x Hitachi 7K80 SATAII :D

I'd run 6x 7K80 id ya had 6x sataII ports :)

Quote:

Originally Posted by Prometheus

The ATI SB450 actually has BETTER IDE and SATA performance than the nVidia nForce4. Check my results at http://www.anandtech.com/cpuchipsets...spx?i=2542&p=9. The only real issues is USB, which is less than half the throughput of nVidia. Sata2 and NCQ are not there on SB450 and I totally agree they should be there to compete with nForc4.

I was merely pointing out that performance wise, SATA2 3.0Gb/s and NCQ really make no performance difference right now. SATA sustained transfer rates are still maxing out at around 100MBps (0.8Gbps), and that's only with RAID-0 and two drives... so SATA1 can easily handle current real transfer rates with it's 1.5Gb/s. NCQ is similarly not very important at present.

It's like nVidia's new Dual x16 SLI. We're all testing it but we can't find any test results that show it makes any performance difference at all over dual x8, but we're still looking.

You ran SATA150 drives on SATA2 port ? Why not run SATA2 drives across all 4 controllers ? time limits of the review ?

Prom maybe i need a copy of Business Winstone 2004 and MCC Winstone 2004 for comparable I/O benchmarks heheh

In Australia
74GB WD Raptor 10k = AUD$259-295

80GB Hitachi 7K80 SATAII = AUD$79-110

4x 7K80 = AUD$320-440 still cheaper than 2x 74GB WD Raptors!
10-19-2005, 08:01 PM
LowRun

Quote:

Originally Posted by MickeyMouse

lmfao I run to this day 3 original dfi nf4's with the bios 1-25 or even the predated beta that shipped on all org nf4 dfi's and it is one of the most stable bios's I have run.

I have a bios 1-21, 1-23, 1-25, 1-26 and the 2-01, lol

So this CPU killer sounds like maybe your PSU or some other Hardware you had was not very user freindly. :)

EDIT: flame removed
-saaya

Yeah, some people just do everything better than everyone :rolleyes:
10-19-2005, 08:06 PM
fareastgq

personally, I liked the graphics card comparision, it gave me hope for the single card solution which I'll be getting, well my stock ECS went to heaven too,,,,.... I guess for me it'll be the expert or the single card solution from ati/dfi
10-20-2005, 03:59 AM
Gambit_2K

Quote:

Originally Posted by Ugly n Grey

The problem with the SATA numbers game is people somehow believe that the bigger number makes a difference. Sustained write performance never enters the equation. The actual internal transfer speed to interface of the drive is the choke point and only the cache bursts at anything resembling those numbers. But 3.0 is better than 1.5 right? Must be, it's a bigger number after all.

Would you rather have a 250hp motor in a 2000lb car or a 350hp motor in a 3500 lb car if your goal is speed?.... Bigger is not always better. The numbers are decieving...

/rant mode

UnG

The REAL problem as I see it is that people on this board believe in what you just wrote. And that's really silly when people on XS are supposed to be enthusiasts and KNOW better then that.

The lack of SATA2 support is not any REAL drawback!

I think ATI should tell the mobomakers to market the boards something like this "ATI IDE and SATA speeds exceed nVidia IDE and SATA speeds even without support for SATA2" :D
10-20-2005, 04:50 AM
4rory

I thought I read somewhere that crossfire will only run if the chipset is compatiable. like some info needs to go threw it.
So that means the graphics drivers check what motherboard you have.

Sata 2 is marketeing, atlest right now it is. kinda like 64bit processors.
10-20-2005, 05:13 AM
cantankerous

Quote:

Originally Posted by Tony

The NF4 ultra D will never run 4dimms 1t, it needs a hardware change to do it.

i will have a bios that does the REAL cas1.5 on the NF4....the one thats there now is actually cas2 it seems.

Hi Tony, when do you expect to have this bios exactly and will it be available to the general public? Most low latency ram users are using CAS 1.5 even if it isn't doing much however would probably like to see what 'true' 1.5 can do if your system is still stable. I know I would like to see any sort of performance increase I can get even if minor.

I am LOVING your 704 mod bios. I am sure the CAS 1.5 bios will be something different on its own as it wasn't edited like you did to the 704 bios however if you could get a CAS 1.5 working bios with all the same tweaks you did to 704 would be simply amazing.
10-20-2005, 06:44 AM
saaya

Quote:

Originally Posted by Prometheus

I was merely pointing out that performance wise, SATA2 3.0Gb/s and NCQ really make no performance difference right now.

true, but its always leaving a bad taste on my mouth going for hardware wich might be less future proof.
i think it was a mistake to go for the slower southbridge, even if it was a lot cheaper or anything. hmmm maybe you can run some tests on this as well? :D

real world impact of the slower southbridge.
game level loading times and boot times as well as copy times for the hdd performence and copying mp3s to an ipod for example as a usb test.
i think this is what people would like to see more instead of synthetical and abstract benchmark numbers. :)

Quote:

Originally Posted by Ugly n Grey

The problem with the SATA numbers game is people somehow believe that the bigger number makes a difference. Sustained write performance never enters the equation. The actual internal transfer speed to interface of the drive is the choke point and only the cache bursts at anything resembling those numbers. But 3.0 is better than 1.5 right? Must be, it's a bigger number after all.

Would you rather have a 250hp motor in a 2000lb car or a 350hp motor in a 3500 lb car if your goal is speed?.... Bigger is not always better. The numbers are decieving...

/rant mode

UnG

exactly...

damn gerge, those numbers are insane!
and they are cheap too compared for the performence you get!
very nice find!

Quote:

Originally Posted by 4rory

I thought I read somewhere that crossfire will only run if the chipset is compatiable. like some info needs to go threw it.
So that means the graphics drivers check what motherboard you have.

Sata 2 is marketeing, atlest right now it is. kinda like 64bit processors.

ive been told crossfire works on nf4 and vias dual graphics chipset at cebit :D
i dont think they changed crossfire a lot since then

64 bit cpus arent marketing, for the desktop, pretty much yeah, but for workstations and servers def not.
10-20-2005, 07:34 AM
eva2000

Quote:

Originally Posted by saaya

damn gerge, those numbers are insane!
and they are cheap too compared for the performence you get!
very nice find!

yes those numbers are insanely awesome - that's why i wish to see the same if not better on a crossfire board :)
10-20-2005, 09:49 AM
Tony

Quote:

Originally Posted by MickeyMouse

lmfao I run to this day 3 original dfi nf4's with the bios 1-25 or even the predated beta that shipped on all org nf4 dfi's and it is one of the most stable bios's I have run.

I have a bios 1-21, 1-23, 1-25, 1-26 and the 2-01, lol

So this CPU killer sounds like maybe your PSU or some other Hardware you had was not very user freindly. :)

EDIT: flame removed
-saaya

How about the 1-14, 1-15, 1-18, 1-19, do you have those? I think not! I do though.

If you decide to run any of those and leave Vcore on auto you stand a chance of pumping 3V + into the CPU which instantly kills the memory controller...

Jason, stick to being the uber overclocker you think you are and i will stick to helping make the boards you dearly love.
10-20-2005, 10:37 AM
Weee

These boards are down to $209 on newegg already? Sweet in 3 weeks they will be $5

Show 100 post(s) from this thread on one page