Better Video Cards Do Make a Difference
I have a HD 3400 class video card and a HD 3800 video card. Both are two generation old ATI cards, but both meet the recommended specs that tech support lists. With most non-game software I have never noticed that better video cards made much difference, but with L4 the HD 3800 makes a dramatic difference, L4 seems almost snappy with it! Granted it does not help with HDD or internet limited functions such as syncing or searching for commentaries in the PG, but anything that requires the displaying of information is very much improved. Right clicking is very close to L3 speeds, the only slowdown is caused by the program looking for information to display, not the displaying of information itself.
The good news is that HD 3800 class cards are selling for around $75 at newegg and there might be some HD4000 class or Nividia cards that are even faster at this price range.
I hope this information is helpful to anyone looking to improve the user experience with L4.
Is it GPU speed or video memory capacity/speed that helps? I say that because the amount of video memory (64 MB DDR to 512 MB DDR2 on a HD3650 AGP) made a huge difference to startup time in Win 7 on a 6 year old machine - at least that seemed to provide the difference rather than GPU/core speed.
Dave
===
Windows 11 & Android 13
Dave,
Both the HD3400 and the HD3800 have 512 MB, however the HD3400 uses DDR2 and the HD3800 uses DDR4. So it is not the size of memory, but the speed of the GPU and the memory that must have made the difference for me. Your HD3650 is certainly faster than my HD3400 and the added memory (64 to 512) and faster memory speed (DDR vs. DDR2) was why you saw an increase in speed.
Video card purchases are a hard call, I typically try to go three years or more with my cards, so I want them to last awhile. On the other hand, I don't like the high markup on the latest and the greatest. If I was in the market for a new card I would be looking at the HD5700 series. They are almost as fast as the last generation HD4800 series and they have Direct 11 support, so they have some good life in them. There are a lot of good deals right now around the $100 range, but for $50 more you would have Direct 11 support and a card that is very energy efficent so it will run cooler and quiter. Three years of not having to listen to fan noise has to be worth something. [:)]
I have not looked at the specs for the two cards; though, I suspect the number of GPUs makes a big difference.
As far as 64MB DDR verses 512 MB DDR2; not only is there more memory, the memory speed is faster, this will make a difference.
Other things that make a difference are multithreaded processors; I see no delays on my XEON, my fast Dual-Core is much slower running Logos 4 because I designed it for single threaded gaming. (Not a fare comparison, the XEON Quad Core with Hyper-threading and an L3 cache is smoking fast)
My gaming machine boots to the OS in less than 30 seconds and opens Logos 4 in about 3 minutes; though, my XEON with Hyper-Threading boots slower, it starts Logos 4 in about 10-15 seconds while I am compiling, running virtual machines ... etc., the XEON keeps up with all the threads Windows can throw at it.
Basically, I would suspect number of GPUs make a difference, machine and video memory size and speed, multi-threading and the cpu cache.
That said, I would expect performance would be as follows (best to least): XEON, i7, Quad-Core, Dual-Core then video, memory ... etc. adding to the performance.
From what I have seen, it looks like Logos 4 performance is directly proportional to how many threads the processor can handle.
Coreluv
Ken,
Some interesting observations. I too am starting to think the L3 cache plays a big role. My church is a Phenom 9660 and I can definately see that Logos 4 is using all four cores. But this first generation Phenom does not have the 1MB L3 that the second generation Phenoms do. My home computer has an Athlon 2 chip, which is basically a Phenom II with two cores disabled. It has the large L3 cache and in some things (especially repeated actions) it is faster than my quad core. I am guessing that this is because it is fetching data from the L3 cache.
The bottom line is if you are in the market for a new computer you want to purchase a well balanced system. With some software the extra cores and caches, good video cards did not make that much difference, but with Logos 4 it seems to use everything you can throw at it, which is a good thing in my mind. I hate purchasing hardware that is rarely if ever used by the software you run. For years hyper-threading and multiple cores have been under utilized by software.
Interesting, it looks like the cache makes a difference - the Quad-Cores actually have more L2 cache, this should also help.
It has just been the past couple of years we started writing SW to take advantage of multi-threading and multiple cores, it is only going to get better now that we have some momentum in this direction.
Coreluv
Or you could take the view that software producers realized that most 'serious' users (i.e. non gamers) would have an average spec PC and coded accordingly.
I would guess Logos3 uses 'ordinary' Windows functions to write to the screen and Logos4 is doing something different...
The problem with all this leading edge technology use is that Logos 4 is still, essentially, a text based research tool and I'm not sure if the functionality it has requires a leading edge approach. Actually, as a C++ programmer who's written this sort of thing for years, I know it doesn't...
I would 'sacrifice' a snazzy look and feel for usability and performance without blinking...
Not that I'm saying that software shouldn't make use of multiple cores and the like, especially as they are becoming more and more common - but maybe it's too early to rely on their presence. And as for a program like Logos4 needing a higher end video card? I don't think I would call that progress, myself.