@Avinash, thank you for bringing more information to the discussion. I look forward to seeing our complaints get crushed by real improvements.

Hello everyone,
First off, I want to take a moment to sincerely apologize for not sharing regular updates with you regarding Illustrator’s performance enhancements. I completely understand the frustrations many of you have expressed about our progress (or lack thereof) in making Illustrator a more multithreaded and multiprocessor-friendly application. Your feedback is vital, and it’s clear that we haven’t done a great job of keeping the lines of communication open.

I want to assure you all that your voices have been heard loud and clear. We are working on making Illustrator multithreaded and multiprocessor-capable. It's a big undertaking, but it's happening! We genuinely appreciate your passion and ideas about how we can improve Illustrator’s performance, and I will ensure that we keep you updated on our progress moving forward. You deserve to know where we stand and what we’re working on.

I know actions speak louder than words, so I'm looking forward to showing you the improvements we're making. Thanks for sticking with us, and for pushing us to do better.
We've been digging deep into the app's architecture, and I want to give you a peek at what we've been up to. We've identified several tasks that were bogging down the main thread, and I'm happy to report we're making progress in moving these to background threads. Some of the areas we have tackled so far: Crash Recovery, Snapping, Thumbnail generation. These processes are now running on separate background threads, which should start to yield noticeable performance gains. We will be continuing to identify and shift more processes off the main thread.

Additionally, we've been focusing on optimizing key areas that users have flagged as pain points. We've made significant improvements in: Image handling, File saving, Applying effects.

Very soon you will see these changes in action. I will share an update once these are in Beta build for you to experience these improvements firsthand. We know there's more work to be done, and we're committed to continuously improving Illustrator's speed and efficiency.

I will continue to update this thread with ongoing progress. I’d also like to invite you to be part of our Customer Advocacy Board (CAB). This CAB will allow for closer engagement, giving you a seat at the table as we explore ways to enhance our product. If you're interested in contributing and being part of the conversation, please let us know through this link: https://survey.adobe.com/jfe/form/SV_czQpFnmd3PWAreS

We appreciate your patience and feedback.

Image traced raster images are a big pain point. I've noticed a number of customer files we have performance issues with include very complex vectorized images (likely scaled down from a larger raster image). Even worse when those are used to make a pattern. Performance is also affected if they're hidden behind a clipping path. Often best to rasterize, but that also takes quite some time. I uploaded an example file to the survey last week. It was a small example around 20mb, but we've had nightmare AI files of over 1GB using complex vectorized images.

The worst offender is stock vector "watercolor" images that are probably just image traced rasters with no optimization at all. They just trace it, save it and good luck opening that file.

That being said those files are working OK on affinity designer not great, but manageable at least, so yeah it can work better, if the application is good.

@Seth I know exactly what you mean. Sometimes I get massive raster files that have to be traced and manually cleaned up, then I have to use the shape builder to optimize it. It can get to the point where it takes a minute or two between operations, just waiting for Illustrator to do something. The final file is usually just a few MB, but you wouldn't know it from the way the program responds to future edits. By then it is 100% vector, but still as slow as molasses. I haven't had a crash yet, from that type of scenario, but I know it is only a matter of time.

Oh, speaking of files you can't upload, I once had to handle a monstrosity that was an image trace of a print resolution wood texture behind a logo. It was something like 48"x36" and was well over a gig and took something like 5 minutes just to open and draw all the layers (there were only a couple). And if you wanted to do anything... you... just... had... ... to... ... w---.... be..... pati-...ent.

That one made Illustrator crash more than once. lol, it was a nightmare.

I imagine Adobe has the means to reproduce such a nonsensical real-world nightmare. But yeah, that's one of the reasons I think Illustrator needs a few more cylinders under the hood, because we don't always get to choose what we work on and sometimes a client (or corporate) hands you a "situation" and you just have to deal with it.

I would have loved to have uploaded a problem file in the survey, but they are all property of my clients, and I do not have permission to share them.

@Neff oh I see, you mean you work with them in a collaboration and don't provide final files. Tough luck, I work with smaller clients and the only time I need to share files are with the printshops that most require PDF anyway.

1.关于填写调查表，我想大家不愿意填写的原因其实很简单， 就是Illustrator团队不够重视客户的意见，提出的建议可能需要等几年后才会开始解决。目前在论坛中的很多小问题、与日常操作息息相关的问题也不能及时得到解决。
2.长期以来Illustrator卡顿、崩溃、缓慢问题一直都没得到有效改善。
3.希望Illustrator团队接下来一方面尽全力优化性能；一方面解决日常使用（甚至是基础的）息息相关的功能，比如：选择功能，画板操作，统一变换，对齐操作，在不同输入法状态的快捷键问题，等等。
4.与Illustrator相比，photoshop做出了很多高质量的升级：启动速度、移动画布速度、滤镜渲染速度等等都非常快，最近推出的“悬停显示图层边界”功能，“自动设置默认字体大小”等功能都非常实用，解决了日常操作的很多痛点。这些都是值得Illustrator学习的。

@nikolas-karampelas that is nice for you, but when my clients (that are also bigger companies) expect me to handle files in AI or PSD because they internally work with that, I can't just tell them to accept PDFs.

@Neff just handle out editable PDFs, I don't know which type of client is expecting ai and psd but I don't provide those files, whatever I use I hand out editable PDFs and jpeg/png.
The safest option and nobody complain to me.

@stuart-chesters In my last tryouts with Affinity, it was in general at least 10 times faster than any Adobe product. The problem still persists in a lack of advanced features compared to Adobe and of course Adobe being the industry standard (monopoly). So when a client comes to you he expects everything to be handled in AI and PSD files, which Affinity doesnt fully support. This is by design from Adobe, as PSD and AI are closed formats, that others need to reverse engineer.

This thread makes me chuckle, as it reminds me of herding cats or as a Yorkshire friend of mine says, 'It's like pushing fog up a ditch with a rake".

That said, I'm glad that we keep on reminding those with the power to make a difference that their product stills falls woefully short and we care about that.

The point I wanted to make was I was genuinely excited as I read through @Seth comment from July 5th outlining what should be possible and a way forward. If only. How great it was to dream for a while about that parallel world where Illustrator got the developer love and dollars it deserves.

I'm also wondering in a genuine attempt to explore alternative solutions, what experience others have had in trying to jump ship to Affinity Designer?

Wishing this community patience, reliable recovery of illustrator crashes and ingenuity in finding work arounds. We're creatives after all!

It is still horribly slow and laggy for complex illustrations. I have noticed no real improvements in any capacity. Again, I don't need Illustrator to save more power for other tasks. I need it to be responsive, quick, and smooth. I don't think I can fill out any more Adobe surveys without swearing. Fix. The. Performance. Issues. That's all we want.

Is this what they call in the US "Gaslighting"?
I believe that Adobe's company vision has recently shifted. It's pretty unfortunate, to be honest.

I'm pretty sure I can speak for everyone when I say this new performance update is yet just another bandaid you guys slapped on a dying pig and is not what everyone is asking for.

The problem is in all your products btw. I'm using Premiere right now and previews in 1/4 with high quality turned off with minimal fx, and only using 10% cpu. 3% gpu, 25% memory, is just a nightmare. It's impossible to get a preview in real time without wasting hours. The issue is not addressing this years ago has made it even harder for you to fix because you buried your legacy code in garbage features nobody asked for. Now you have a giant mess that is virtually impossible to tear down and rebuild.

It's too late for Adobe. It would take less time for a competitor to build a new product from scratch while using Adobe as a cautionary tale to avoid the big mistakes.

Juste installed and tested 28.7: When loading a bit heavier files the interface still freezes and the top menue still fades out to white till AI catches itself again. Exactly like in the current stable release.

Zoom indeed is a bit better (but still far from good - or even smooth). But when I move graphics or vector elements, they still DISSAPEAR which makes them nearly impossible top place. In GPU view I even don't get the vector outlines. Nothing here which blows me away, but honestly didn't expect anything else. So sad, Adobe, so sad...

Zoom is still absolutely lame and universes behind what's possible with modern computer systems. I've an NVIDIA RTX4060 Ti 16GB which empowers most of my software a lot... but Adobe. TopazLabs AI scaling/repairing/enhancing runs ultra fast, half of the times in a blink of an eye (near realtime). Can't afford to use Adobe products here, because I haven't that much time... Pan and zoom in Blender works far superior - in 3D instead of simple 2D Illustrator...

I wanted to add a zip to your survey as you ask for example projects there - but your upload limit is 50 MB. That was 450 too little...

Adobe, this is not "done" at all. You've barely even started. This idea and the support behind it wasn't for a token acknowledgement that multicore systems exist and a surface level of support in only a few aspects of the application.

I'm really curious where you get "up to 5X faster" from, particularly given what core counts are like on most enthusiast and pro-level machines these days. Even mainstream CPU's these days have 10+ cores. Even many laptops offer more cores. So where does such a miniscule number like "up to 5x" even come from? We're expecting _actual_ multithreading, not just optimizing a single function of Illustrator here and there. But don't take that for criticism, optimization when it doesn't break things is a net positive for customers and customer satisfaction.

And considering pan and zoom should be handled by the GPU, the improvement should be an order of magnitude greater. Core counts on even budget GPU's are significantly higher than 10. I just looked at a $50 GT 710 that nobody in their right mind would buy. It has 192 CUDA cores. ONE HUNDRED and NINETY TWO. Not ten.

Even if you were using other cores for others tasks (which you should be) there's PLENTY of room for more than 10x improvement. Considering most of us are probably using primarily vector art, the processing should be insanely simple to compute. The bread and butter of GPUs you might say.

Here are some areas for potential improvement which should be relatively low-hanging fruit:

1. Processing Appearance Effects (adding extra fills, strokes, transforms, shadows, etc. and blend modes)
#1 is probably the area with the most significant potential for improvement in my experience. And honestly, where so much of the complications come from. This absolutely should be the focus of optimization and stability for Illustrator because so much "power" derives from this seemingly minor facet of the program.

2. Vector Operations like shape booleans, path simplification, etc. This, to me, should easily be the undisputed fastest aspect of Illustrator. Vectors should be unquestionably stable, fast, flexible, and accurate. I expect a GPU focus for vector math would provide incredible speed benefits. It might even be the case that the current functions don't take advantage of newer CPU instruction sets that could also provide speed and efficiency benefits.

Pan, at the very least, should meet or exceed 60fps at ALL TIMES, in any condition. I wouldn't expect Zoom to perform quite as well because it requires recalculation, but Pan is just translation; there should be absolutely minimal overhead.

Additionally, while accelerations and optimizations are made to things like appearance processing, I think redrawing and recalculating should be offloaded to a non-blocking background thread that is not fatally tied to the main process. I expect this is where a lot of errors and crashes occur and it should not be possible for the main process to crash from something this simple.

There should probably be a little progress bar in the appearance panel that indicates if it is processing in the background. An outline or ghost of the object should indicate it is being worked on (should be user preference, some might want the last known good state to simple stay there until processing is done).

In the event there is a failure, there should be a warning flag raised in the appearance panel. The warning should be non-blocking and non-modal unless the user wants that, but it should absolutely not be fatal to the main process. Perhaps an overlay flag could also be pinned near the offending element(s) to indicate an error status to the user.

A background tasks panel should probably also be available to indicate what Illustrator is working on and perhaps also highlight processes that are taking longer than expected. Having background, parallel, non-blocking, managed threads for processing would also allow recovery from otherwise fatal errors by allowing error handling to monitor the threads and catch exceptions before the main process gets nuked in the collateral damage, losing work, state, etc.

For background processing of things like appearance, users might also appreciate "progressive" or "scanline" render previews. This would provide immediate visual feedback of progress, and provide an interruptible indication of final state in case the user wants to tweak something before waiting for the final result, like maybe a shadow isn't the right intensity, blur, distance, etc. Rather than waiting for the entire process, being able to see and interrupt would also probably help users by reducing frustration and wasting time.

(wow, that got long.)

"Hey, these stupid user are STILL calling for Multi CPU support."

"Oh, just tell them we make some performance improvements and mark the topic as resolved."

Is this a joke Adobe? Can I in the future just pay a part of my subscription fee. when I say your demand was resolved by sending you 10 bucks?