• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

HD5770x2!

Status
Not open for further replies.
It will take some time for that "eye inffinit" multi display to catch on IMO. Especially with the way the economy is. I look forward to Brent's HD5770 review.

I don't consider myself a fanboi. I'm just having fun with my toys and enthusiastic about the knowledge that I have.

Right well I already have 3 displays so for me it starts as soon as the 2gig cards hit the market.

But anyway a wise man once said knowledge is power but ignorance is bliss. So if you got it you might as well enjoy it.:cool:
 
Uh, duh.
I think that is what we all have been telling you for 14 pages.
Thank you for finally seeing the light. Can you tell your buddy that please.
We won, we won. The war is over, hip hip hooray!

I still think the HD5870 would be a more optimium peformance choice over a HD5770x2 even though it would cost more.

The HD5770x2 is not designed to beat the HD5870. HD5850 yes.

If the HD5770x2 wrere to launch tommorrow I wouldn't expect it to cost less than $299.99 with the way the HD5000 series cards are priced right now.
 
Explain to me how 3 out of 6 games is most games?......

Because it actually wins in crysis :rolleyes: and in Fallout its got a higher average but a lower max and min, so thats at least a tie.
 
Right well I already have 3 displays so for me it starts as soon as the 2gig cards hit the market.

But anyway a wise man once said knowledge is power but ignorance is bliss. So if you got it you might as well enjoy it.:cool:

If you have the money enjoy it.
I'm going to take it easy on the spending till the economy recovers.
 
No plank, it would not. It would cost more to manufacture and not perform better the a top end single GPU. When they do make an X2 it won't use a crippled 5770 but a pair of 5870's and will be an uber part. When it comes out and at what price depends on our good friends at the green factory. So far it looks like the red camp has no need to rush anything. Especially low end X2 parts that cannibalize their own sales.
 
Still looks to me that the amount of memory (512mb) is holding it back, an X2 card would need to have 2GB (1 per card) and it would help a lot. Even in that review the 4770 is beating it by a fair margin in MOST of the games.
yeah but the bus width is an issue too because it doesnt get doubled either even though you think it does.

"The HD 4770 CrossFire configuration’s memory bus is also limited at 128-bits wide" http://www.hardocp.com/article/2009/08/18/amd_radeon_hd_4770_crossfire_evaluation/10

now please stop claiming that crossfire doubles the memory bus width because it doesnt.
 
It's going to take a special HD4850 to beat my HD 4770 clocked at 960/1075.

HD 4770x2 would be interesting is SFF systems. With the power consumption budget of one HD 4850 you could get much higher performance.

Out of curiosity, does the 4770 get to 960 on stock voltage? It would take 768 on the core of a 4850 to match that, but the voltage would have to be modified I'm sure. 1075 RAM on the 4850 is the same bandwidth as 1075 on the 4770.
 
Btw I used both the crossfire HD 4850 and HD 4770 setup.

With Ocing the HD 4770 setup was faster and ran a LOT cooler so was FAR easier to get silent.
 
Out of curiosity, does the 4770 get to 960 on stock voltage? It would take 768 on the core of a 4850 to match that, but the voltage would have to be modified I'm sure. 1075 RAM on the 4850 is the same bandwidth as 1075 on the 4770.

No it was at 1.2 volts, however all original reference HD 4770 cards had programmable voltage circuits :D

Combine that with the fact the 4850 is already a furnace, I would not want to over volt it.
 
Out of curiosity, does the 4770 get to 960 on stock voltage? It would take 768 on the core of a 4850 to match that, but the voltage would have to be modified I'm sure. 1075 RAM on the 4850 is the same bandwidth as 1075 on the 4770.

Software voltage mod.

Very easy to do. Nothing like doing a hard voltage mod to a HD4850.
 
I still think the HD5870 would be a more optimium peformance choice over a HD5770x2 even though it would cost more.

The HD5770x2 is not designed to beat the HD5870. HD5850 yes.

If the HD5770x2 wrere to launch tommorrow I wouldn't expect it to cost less than $299.99 with the way the HD5000 series cards are priced right now.
it would not beat the 5850 is most cases either because at settings that people would need that much gpu power the 128bit bus would hold it back. the 5850 would beat the 5770x2 in almost every realistic gaming scenario while costing cheaper.
 
HD4770 oced @1000/1150 run faster than an hd4850gs oced @750/1100!

sapphire hd4770@1000/1150:
18676-2.jpg


gainward HD4850GS GDDR3 @750/1100:
15851.jpg



 
Because it actually wins in crysis :rolleyes: and in Fallout its got a higher average but a lower max and min, so thats at least a tie.

did you even look at the review?

Look at resolutions and AA levels...

Man you post on [H]ardocp and dont even understand their reviews

/facepalm
 
I still think the HD5870 would be a more optimium peformance choice over a HD5770x2 even though it would cost more.

The HD5770x2 is not designed to beat the HD5870. HD5850 yes.

If the HD5770x2 wrere to launch tommorrow I wouldn't expect it to cost less than $299.99 with the way the HD5000 series cards are priced right now.

So close yet so far away. To make the 5770X2 would easily cost more. Why you would think otherwise I can't even guess. Lets start with double the DDR5 cost as a starting point. And again, we don't care what price YOU think a 5770X2 should cost, it would cost more to make one then a 5870. So why would ATi price it lower, and lower the sales of the 5870?
Wanna job at Nvidia?
 
yeah but the bus width is an issue too because it doesnt get doubled either even though you think it does.

"The HD 4770 CrossFire configuration’s memory bus is also limited at 128-bits wide" http://www.hardocp.com/article/2009/08/18/amd_radeon_hd_4770_crossfire_evaluation/10

now please stop claiming that crossfire doubles the memory bus width because it doesnt.

Ok if the memory bandwidth was not affected in any way then surely the scaling to the crossfire setups would be much worse than it is. Each card has its own frame in AFR so while they store the same data they can both read off it at the same time, its effectivly doubling it.
 
Ok if the memory bandwidth was not affected in any way then surely the scaling to the crossfire setups would be much worse than it is. Each card has its own frame in AFR so while they store the same data they can both read off it at the same time, its effectivly doubling it.

Just thought I would give you a heads up since you dont understand [H]'s reviews

"Looking at the actual gameplay experience, we find that the Radeon HD 4890 provided the best as a whole. Arma II was better experienced on the Radeon HD 4890, so was Fallout 3 and Crysis: Warhead. Every other game tested exhibited the same gameplay experience between the Radeon HD 4770 CrossFire and Radeon HD 4890."
 
Sigh, now your relying on subjective measurements that go against the test data?

Hell when I played 2x HD 4870 vs OCed 2x HD 4770, I would be damned if I could tell the difference.

I have owned and tested both and IMO the HD 4770 were better simply due to their ease of use and ability to stay quiet.
 
Ok if the memory bandwidth was not affected in any way then surely the scaling to the crossfire setups would be much worse than it is. Each card has its own frame in AFR so while they store the same data they can both read off it at the same time, its effectivly doubling it.
then why does not one single review of a crossfire or sli setup of any sort say anything like that? maybe you should inform these reviewers because anytime they mention bandwidth they say its the same. like I said earlier it was one of the gtx295 cards criticisms. it did not have the bandwidth of even the gtx285.
 
HD4770 oced @1000/1150 run faster than an hd4850gs oced @750/1100!

sapphire hd4770@1000/1150:
18676-2.jpg


gainward HD4850GS GDDR3 @750/1100:
15851.jpg





Hmmm Lets see some Game benchmarks. Synthetics means nothing "Now thats thinkin with your dipstick jimmy!"
 
Sigh, now your relying on subjective measurements that go against the test data?

Hell when I played 2x HD 4870 vs OCed 2x HD 4770, I would be damned if I could tell the difference.

I have owned and tested both and IMO the HD 4770 were better simply due to their ease of use and ability to stay quiet.

I like what you said in rev 2.0 best
 
The HD 4770 is a bandwidth limited card at stock. Right?

Then how come adding another card unless crossfire is somehow mucking up does not give a 0% performance increase?

In fact the HD 4770 scales really damn well.
 
Hmmm Lets see some Game benchmarks. Synthetics means nothing "Now thats thinkin with your dipstick jimmy!"

Why does this have to be an argument every single time? :confused:

His benchmark against any other card's benchmark is all that is required to understand the performance measured.
 
I like what you said in rev 2.0 best

I would like to point out I am basing my views on having OWNED the following setups:

2x HD 4850 512mb
1x HD 4850X2 2GB
2x HD 4870 1GB
2x HD 4770 512mb

While testing the dual cards individually aswell. IMO the 4770 was the easiest to OC in crossfire, provided very similar results to my 2xHD 4870 setup and did while silent with a couple of £16 coolers.
 
Man, I've seen plenty of arguments when talking about Nvidia vs. ATI, but I guess the new thing is ATI vs. ATI now. The 4770 is a good card, but as someone else has already stated, the 5770 will not be in the same position compared to the 58xx series as the 4770 is against the 48xx series. They are ALL 40nm. :D
 
The HD 4770 is a bandwidth limited card at stock. Right?

Then how come adding another card unless crossfire is somehow mucking up does not give a 0% performance increase?

In fact the HD 4770 scales really damn well.
a single 4770 wasnt really bandwidth limited. having two of them was and thats one of the reasons that the 4890 could run better settings and still beat the 4770 crossfire setup in some cases.
 
Why does this have to be an argument every single time? :confused:

His benchmark against any other card's benchmark is all that is required to understand the performance measured.

I'll let [H]ardocp explain it to ya then

"Looking at performance, it is a mixed bag as we have discovered. In shader intensive games the Radeon HD 4770 CrossFire is going to give the HD 4890 a run for its money. In memory limited games or games that allow high resolutions with AA settings, and high detailed textures, the HD 4890 is going to perform better. If, however, you are running specialized applications (i.e. not games) such as Folding@Home, or other computational oriented applications, then Radeon HD 4770 CrossFire is going to provide greater performance for you over a single HD 4890."

Games > 3dmark
 
I would like to point out I am basing my views on having OWNED the following setups:

2x HD 4850 512mb
1x HD 4850X2 2GB
2x HD 4870 1GB
2x HD 4770 512mb

While testing the dual cards individually aswell. IMO the 4770 was the easiest to OC in crossfire, provided very similar results to my 2xHD 4870 setup and did while silent with a couple of £16 coolers.

No thats not what I mean. You revised the post I commented on three times. I read each and thought rev 2 was the one I think you should have gone with.
 
I'll let [H]ardocp explain it to ya then

"Looking at performance, it is a mixed bag as we have discovered. In shader intensive games the Radeon HD 4770 CrossFire is going to give the HD 4890 a run for its money. In memory limited games or games that allow high resolutions with AA settings, and high detailed textures, the HD 4890 is going to perform better. If, however, you are running specialized applications (i.e. not games) such as Folding@Home, or other computational oriented applications, then Radeon HD 4770 CrossFire is going to provide greater performance for you over a single HD 4890."

:confused: Again apples to apples just compare the benchmark. It's so easy to do. You seem to know that the HD4770 is a great card but don't want to admit it.
 
:confused: Again apples to apples just compare the benchmark. It's so easy to do. You seem to know that the HD4770 is a great card but don't want to admit it.

A benchmark of what? a perty 3d image going across the screen to look at?

I am sorry I want real games that I can play and accually do something in.

Synthetic benchmarks mean nothing bro.

Great card, but a 4980 > 4770 in crossfire and 4770x2

And incase you ask me to show you proof

http://www.hardocp.com/article/2009/08/18/amd_radeon_hd_4770_crossfire_evaluation/1

Right here at [H]
 
:confused: Again apples to apples just compare the benchmark. It's so easy to do. You seem to know that the HD4770 is a great card but don't want to admit it.

That isn't the topic! Quit trying to derail your own thread. You can't pull a rabbit like that. You were cheer leading why a 5770X2 is a great idea. There have been many good responses why not, when are you going to answer any of them? I mean, nut up dude. Why do you think it is a good idea for ATi to deliver this product?
Come on, one solid answer. You can do it!
 
Great card, but a 4980 > 4770 in crossfire and 4770x2

Single overclocked HD4890 in my RIG was no match for 2 HD4770s Xfired in gaming.



Take a look at the world record benchmarks for a single HD4890 and a mulit HD4770 set up. You will see that 2 HD4770s destroy a single HD4890.


http://hwbot.org/hardware/videocard/radeon_hd_4890

You want game reviews do a search on Google for HD4770 Xfire vs. HD4890.

There is alot of information out there that proves 2 HD4770s in Xfire are more powerful than a single HD4890.


HD4770 is an important building block to the HD5770.
 
Last edited:
Single overclocked HD4890 in my RIG was no match for 2 HD4770s Xfired in gaming.



Take a look at the wolrd record benchmark records for a single HD4890 and a mulit HD4770 set up. You will see that 2 HD4770s destroy a single HD4890.


http://hwbot.org/hardware/videocard/radeon_hd_4890

You want game reviews do a search on Google for HD4770 Xfire vs. HD4890.

There is alot of information out there that proves 2 HD4770s in Xfire are more powerful than a single HD4890.

http://www.hardocp.com/article/2009/08/18/amd_radeon_hd_4770_crossfire_evaluation/

Games > 3dmark Record breaking uselessness

Proof right here on this website man :)
 
That was a good review but the only weakness was that they didn't properly over clock the HD4770.

HD4770 can go much faster than the speeds they tested at.

And so can the 4890 :)

Seen many of them get 1ghz
 
Ah, I get it. It wasn't a 5770 X2 argument all along. It was another 4770 is my girlfriend thread. Why didn't you just say so from the beginning. It would have saved a lot of time and effort. We were hitting you with lots of reasons why 5770X2 is just stupid, but you never seemed to care. I understand now. It's okay. You win.
 
And so can the 4890 :)

Seen many of them get 1ghz

did not see sonda5 meaning an hd4770 gpu@1ghz is running faster than an hd4890 gpu@1ghz!:rolleyes:

Anyone know that an hd4890@1000/1150 outperform an hd4770@1000/1150 same speed but quite different bus speed!
my sapphire hd4890 runs @1000/1200 wit softmod and i did not try to face my hd4770oced with this card at this clocks...
 
Crossfire memory scaling:

AFR vs card twice as fast:

Cards 1 and 2 work together, they can access two letters at a time.
Card 3 is twice as fast and can access four letters at a time.

Letters = required data for frame

Frame 1 = A + B
Frame 2 = B + C
Frame 3 = C + D
Frame 4 = E + F

Over a set period of time:

Crossfire setup:

time 0:
Card 1 accesses A + B renders frame 1
Card 2 accesses B + C renders frame 2
time 1:
Card 1 accesses D + E renders frame 3
Card 2 accesses F + G renders frame 4

Single card setup:
time 0:
Card 3 accesses A + B + B + C renders frame 1 + 2
time 1:
Card 3 accesses D + E + F + G renders rame 3 + 4


Depending on optimisations it may not need to access B twice, depends on the buffers. Lets say it has to call it twice. Over the same time period both cards accessed 4 letters even though each card in the crossfire setup can access 2 letters at a time.

Crossfire = memory read bandwidth increase, how much depends on what optimisations you make.
 
And so can the 4890 :)

Seen many of them get 1ghz

As a percentage that's less, so both at 1ghz the gap will close.

Please remember that the backends are identical. The shader outputs are different, but the pixel and texel output is the same at the same clockspeed.

Also the gap depends on the games you run, 6 games is not many, have a read of this:

http://en.expreview.com/2009/04/28/first-look-at-radeon-hd-4770-crossfire-benchmark.html/4

With overclocking more even that 3/3 split in the hard OCP review could easily change even their subjective "even though it gets more frames we like this one better".
 
Last edited:
Status
Not open for further replies.
Back
Top