4 points
1. Topaz is terrible for putting out versions that are basically betas. On the FB pages there a tons of people complaining the new versions are so bad they can’t use them (freezing, doing nothing, and crashing are the biggest complaints) with pc users having the biggest issues
2. Their code is poorly written which causes the performance issues especially with the pattern matching and machine learning which is what they call AI, which it isn’t. I use a 2011 iMac and have zero performance issues with lots of other apps doing pm and ml such as PhotoShop , LightRoom , On1 , even Luminar 4. I have no better results on a new mbp i7, max ram, and an SSD drive. In fact, often the iMac is faster on same image.
3. Their pricing model is a ripoff. Used to be you bought a product and got free versions for life. That is not sustainable unless you continue to increase your customer base at a rate higher than the revenue you would get from charging upgrades. If you purchased up to about a year ago you are grandfathered into free upgrades for life for THE products you own. It is easy for a company to rebadge a product and claim it is new and not covered by the lifetime upgrade policy. They replaced that with the new policy which is a hybrid of perpetual and SaaS-based pricing. Perpetual because you own whatever product for life and SaaS because have a contractual process that charges you a set fee yearly. No problem, other than the set fee is way too high based on the norm for software. It is a sign of financial issues and a red flag.
4. Ive tried all the versions of topaz and found their results to be sketchy at best with one noticeable exception, Topaz Studio 2. It is not under the draconic pricing model probably because they arent doing on any work on it. It includes AI Clear which often does a better job of both denoising and sharpening than the expensive standalone they sell.

My 2 cents

I tried DeNoise AI and Sharpen AI free for a month - good results but soooo slow. I spoke to their techie and he said that although I have a brand new MacBook Pro my i5 chip doesn't meet their minimum requirements - I would have needed to get the i7 chip or faster. So, I learned how to do noise reduction with optimal sharpening in Affinity Photo and am very satisfied with results. For certain lighting conditions the Topaz AI would yield better results but for 95%+ of my shots my processing is as good.

It's not surprising you are finding it slow as molasses. The Radeon Vega 8 is a lower end integrated graphics chip which has no built in memory of its own, sharing your computer's memory instead. It has around 15-20% of the performance of many typical midrange graphics card today, and about 5 or 6% of the performance of the best cards available. It is highly unlikely that Topaz uses your computers graphics capabilities at all. The level of performance when running Topaz Denoise AI is mostly related to the use of the graphics card's GPU. My desktop's graphics card is an nVidia GeForce GTX 1050 ti, which is more towards the lower end of current midrange cards but still has 3 times the relative performance of your Radeon. Denoise uses my GPU and as a result it performs very well on my machine. You may have already discovered that a lot of graphics heavy software today will struggle on your machine.