Photoshop CS4 is a disaster

[_ProArtist_]

I'm am just at a loss of words.

What a mess. It could not be any slower. What were you thinking Adobe?

You ripped apart the code just to add GPU support for what? To provide worse performance?

Make sure you DL the demo first... CS4 is a disaster.

The latest hardware cant even run it smoothly... Dont tell me its graphic drivers.

[Walter_Elias]

i Are you willing to volunteer yours?

I'll give them mine in return for a new quad core.

[shoot_me_now-d3u3se]

:)

[_Adam_Jerugim_]

See post #48. No need for any of us to fly anywhere...

On Friday I set up an older box (3GHz Pentium D, dual-core) with a clean XP Sp3 and can now reproduce many of the issues reported. We're in the process of investigating and I will provide more information when I have it.

-Adam

[Guest]

you're a gem adam. thank you for all the help and support you've provided for the last several releases.

[CR_Henderson]

>dave milbut:you're a gem adam. thank you for all the help and support you've provided for the last several releases

Ditto...

There is also the fact that he is online providing useful information on Sunday. That shows either dedication or he has no life. I suspect the former for him while for me it's probably the latter :)

[_ProArtist_]

Adam, make sure you test in vista as well (64-bit especially). These performance issues do not only apply to XP3.

Thanks for being on top of things.

[dpick2]

I'm using vista x64 with a Nvidia 8800 vid card, and it works great.

[Farrellart]

same...

vista x64
Asus P5Q deluxe
16gb ram
nVidia 8800 gts 640mb
q9650 3ghz
loads of drives
This system only goes online to activate software / download updates.

CS4 runs really well.

There are so many variables.....hard drives / ram / drivers /etc... any bad combination could cause issues.

For those having problems, I'm wondering if it's an OpenGL and Direct X conflict, in some cases, as they're two separate systems that are trying to achieve the same thing. Maybe a Microsoft DirectX update has interfered with OpenGL in some way.

[Guest]

Adam Jerugim,

I will be following this thread to see what you find.

It sure is an improvement over M-soft's "trouble shooting" which blames BitDefender running under Vista for every problem I have had. That includes those with the cable un-plugged and BitDefender not having been started.

Thanks,
Bill

[Chris Cox]

Actually, Chris keeps saying "update the drivers" because people are mentioning known driver bugs that the video card makers are supposed to be fixing (but obviously haven't yet fixed for every model and OS combination). We need to get past the known driver bugs so we can find any as-of-yet unidentified bugs.

[Guest]

Chris Cox, I go to the ATI site Every Day to check for new drivers. So far, no luck. I have contacted them three times, via email, and they have nothing to say about the issues I'm seeing with CS4. They tell me to re-install the driver package, which I've done about four times.

I read that Adam J. has reproduced the lag problems on a machine there at Adobe. Do you or Adam have any info on that? Are you two on speaking terms? Sounds to me like there might be identified bugs.

[Chris Cox]

Adam's office is next door to me - we talk, well, a lot. There are several "lag" problems being discussed.

But things like "incomplete redraws" and "incomplete brush cursors" are known driver bugs, and updated drivers are the only thing that will fix them. Again, we have to get those known issues out of the way first before we can start looking for other problems.

[Guest]

>updated drivers are the only thing that will fix them.

OK, thanks. Does me no good, though, because ATI won't admit to any problems with their drivers. They care about gamers, not PS users.

If I knew what ATI drivers you used for testing CS4, I would roll back to them. Then maybe CS4 wouldn't be a problem for me.

[Hudechrome-sd9sPI]

Bugs.

Team A develops a program or an app. They are the basic engineering team. Their work becomes the core for Team B. Team B has been created or enabled by marketing to take A's work and put a skin on it, and wouldn't it be nice to ad this feature, make that feature more useful or glitzy, etc. B develops what marketing wants. Now, the design begins to be buggier than the early alpha version of Team A. In fact, when the remaining bugs for A's product gets resolved, team B's product generates evenmore bugs. So they talk. Team B files a bug report that starts out in the setup description with "Install the Team B app". "Hold on!", says A. "We didn't create that. You did." "Yeah", says B, "but when you fixed big number XYZ, we have another problem show up." A says that's not our problem. But in the spirit of cooperation, A takes a look, and, with it's core running fine they cannot generate the bug. Bickering starts. Team A issues a final release, whereby Team B rejects that ....

You know the story, I'm sure. Now, Team A and Team B are separate companies, and the battle is more like a marriage gone bad.

Could it be that Adobe is asking the video card and driver to do things the card vendors haven't anticipated, and in fixing things for Adobe users, it fouls the nest for other users, causing a cascade of bugs and fixes? On both sides?

[_Adam_Jerugim_]

As promised, here's more info:

On the system in my office that I'm seeing the problem on (Pentium D 3GHz, 4GB RAM, XP Sp3), using a combination of the AllowOldGPUS_ON.reg and then using the following settings in the GPU advanced settings:

Vertical Sync: enabled
3D Interaction Acceleration: enabled
Force Bilinear Interpolation: disabled

Advanced Drawing: enabled
Use for Image Display: disabled
Color Matching: disabled

The brush lag, in my case, is mitigated with these settings. Disabling Open GL Drawing altogether brings the brush performance nearly in-line with CS3.

Still investigating...

-Adam

[Guest]

Adam, no joy over here. There's still about the same amount of lag, and if I drag a window the content disappears.

I would add, though, that with AllowOldGPUS_ON.reg in place, 3D Interaction Acceleration is grayed out in the Advanced GPU Settings.

What video card are you using?

[_Adam_Jerugim_]

In my case, I'm using the nVidia 8800 GT (512MB VRAM). It is worth noting that the same card runs CS4 without any issue on three other systems in my office (one XP SP3, two Vista64).

[Walter_Elias]

Adam,

As stated in previous posts, my experience is quite different than what you're experiencing. Disabling OpenGL makes performance worse. Using the settings you listed above also makes it worse. Again, the only way I get decent (but not great) performance is:

AllOldGPUS_ON.reg in place
Enable Open GL
Disable all Advanced settings

Any other combination of settings brings the problem back.

It seems the problem is very hard to pinpoint.

Maybe you should give some us a phone call, ask us to try different settings, and give you instant feedback. I'm ready and willing to spend the time to help you get to the bottom of this.

[Dave_Evanson]

Adam,
When you say the same card is it physically the same card tried in different PCs or do the other PCs just have the same model of card?

I wonder if it could be related to the card bios or even the manufacturing date of the card, perhaps some un-documented changes were made to the card hardware for different production runs.

[Kwan Parker]

Occurs to me that in some instances problems may not be caused by drivers. The presumption is that the card itself is functioning properly.

A year ago my 7900GS was stricken with the infamous Nv4_disp.dll Infinite Loop . . . intermitently. The only consistent advice from users on Nvidia's forums and in general was to update the drivers or revert to an earlier working version.

After several months, under warranty, I tired of looking for a solution and after performing Dell's hardware video tests, replaced the card. End of of problem.

For peace of mind, those experiencing problems might want to make absolutely certain their card does indeed function properly. Just because the card is new does not guarantee it is working as it should.

Just a thought . . .

[Guest]

Although there are a lot of people posting here with genuine problems (and I really sympathise) I wonder what percentage that represents of the total sales of Photoshop CS4 so far.

[Hudechrome-sd9sPI]

Has anyone actually replaced a card with a different version, obtained the most current updates and said problem solved? Without a series of such events, we do not have any basis by which to say the bug(s) are resolved.

In software testing, when we have filed a bone fide bug, it has to be resolved by the engineers then tested over several platforms before QA will close it.

At least, that's the way we do it where I work doing such testing.

I would suggest that Adobe actually have a couple of cards know to be good available for others to test in their systems. It would go a long way to root out exactly where the problem lies. It needs to be the same card version and driver.

The resulting list of systems that were fixed and those it didn't would be highly useful!

[_Paul_Budzik_]

So far, it seems that Adam has been able to duplicate many of the problems. His suggestions so far have only been workarounds that he says will get you "close" to the performance of CS3. It may be driver issues and/or Photoshop bugs, but too many people, including Adam, can demonstrate the problem so I'm betting that it isn't a "defective" card issue. And in my case, I have a $900 card that runs my other 3d apps perfectly, so I'm not that willing to just go out and buy another for testing purposes, although I might if I was the only one having the problems.

Paul

[Hudechrome-sd9sPI]

To each his own, but if Adobe can demonstrate conclusivly that, with a system basicaaly configured to be reasonably close to universal, this card and this driver will perfom as speced by Adobe. In the interests of getting feedback, offering a short time loan to qualified users would, IMO, garner significant additional data points to be valuable.

My opinion anyway!:-)

[Hudechrome-sd9sPI]

Paul, your position on expensive cards is tantamount to expecting a $50,000 Lexus to perform as well in extreme conditions as a 10 year old jeep Wrangler then blaming the terrain for the problem.

There appears to be no universal autos and no universal video cards. :-(