What happened to Alpha Five desktop?

[Jeff Fried]

I am so dismayed to find out that there is a vast difference in performance between V10 and V7 (yes, I'm skipping V8 and V9 for this analysis because V7 is what has been running in the live environment).

The project was to rewrite an order entry application that now includes many bells and whistles. At the same time, computers were upgraded along with the server. The computers started with Vista 64 bit but were upgraded to Windows 7 64 bit. But factor all that out and just focus on the variable V7 vs. V10. In a type of "laboratory" environment I set up with several of my client's computers so I could take my time to investigate why the revised app is not ready for prime time, I studied many combinations.

But in the end, using XP with a vanilla server, I discovered that my client's old application -- the one they currently use in production, and have been using for many years -- running it under A5V10 has a noticeable and substantial slower performance and response than running it under A5V7. I did many double-takes on this. At its basic level, just starting a new order which pulls up a list of customers to choose from, then proceeds with a pulldown of products to select from, is substantially different between V7 and V10. In V7, all is immediate response. In V10, hesitation. Not a big hesitation, but one which, when four or so people are pounding orders in, there will be a major performance hit to their operation. As to the revised application, loaded with many more capabilities, I expect disaster. Unfortunately, I can't yet test the new app in V7 vs V10 since it won't run in V7 (designed in V9, then V10, so refuses to run under V7). I'm thinking of rewriting the main data entry screen using V7 just to be able to test the new app under V7 vs. V10.

I doubt anyone has a solution to this as I have boiled this down to barebones testing. There is definitely, without question, a major decline in Alpha Five desktop from V7 to V10. I've had it with the V10 hype! I pass along the promise of the newer version to my clients, then end up with an unworkable solution in the end. Has Alpha Software's focus on wep apps killed Alpha Five desktop?

Jeff

[Scott_Walker]

Re: What happened to Alpha Five desktop?

Jeff,

Maybe I'm misunderstanding you but I would expect that as a program gets more feature and therefore more overhead, it would run slower on any given set of hardware then the smaller, lighter, less fully featured version from a few years earlier. That makes sense to me. Thus many times upgrading your op sys or application program leads to the need for more horsepower in the hardware.

Regards,

Scott

[Jeff Fried]

Re: What happened to Alpha Five desktop?

Jeff,

Maybe I'm misunderstanding you but I would expect that as a program gets more feature and therefore more overhead, it would run slower on any given set of hardware then the smaller, lighter, less fully featured version from a few years earlier. That makes sense to me. Thus many times upgrading your op sys or application program leads to the need for more horsepower in the hardware.

Regards,

Scott

Yes, you are misunderstanding me.

First, I'm talking about "the smaller, lighter, less fully featured version" being tested between V7 and V10.

Second, I'm talking about testing "the smaller, lighter, less fully featured version" on significantly upgraded computers. The old application does not require much -- it is still running in production mode on older, underpowered computers with totally acceptable performance. Sure, a newer version of Alpha with more features has the potential of running things more slowly, all else being equal. What I'm talking about here is a significant performance decline. And for what I'm talking about, the test is with XP, not a newer OS, as I stated.

Jeff

[martinwcole]

Re: What happened to Alpha Five desktop?

I'm surprised you never went to V8 - which was appreciably faster than V7.

I have not done extensive micro testing (note that Ira Perlow has a utility that will help you measure times for scripts and parts of scripts for tracing where the differences may be), but I recently moved a big app from 8 to 10 and did not notice any difference in performance - running on Server 2003.

In preparing to move my biggest app from V8 to V10, I tested in both, where it rebuilds all the indexes from scratch running through about 360 tables, and the timing was the same, XPPro - but not faster.

[Jeff Fried]

Re: What happened to Alpha Five desktop?

I'm surprised you never went to V8 - which was appreciably faster than V7.

I have not done extensive micro testing (note that Ira Perlow has a utility that will help you measure times for scripts and parts of scripts for tracing where the differences may be), but I recently moved a big app from 8 to 10 and did not notice any difference in performance - running on Server 2003.

In preparing to move my biggest app from V8 to V10, I tested in both, where it rebuilds all the indexes from scratch running through about 360 tables, and the timing was the same, XPPro - but not faster.

I started to look at V8 but didn't go the distance, so based on what you are saying, I'm going to take a closer look at it. I know it was the first A5 compatible with Vista but not sure if it makes a difference with Windows 7.

You obviously have not encountered the same experience as I did with V10 and can't understand why. I wonder if the fact that you run on a domain server vs. the peer-to-peer network they run on makes a difference. Regardless, the same app they've been working with for years on their peer-to-peer network declines in performance once V10 is introduced. Again, if Server 2003 overcomes this performance discrepancy, I need to be sure it will solve the problems before recommending that investment. All their hardware and wiring is now gigabit, so I'm not sure.

Jeff

[martinwcole]

Re: What happened to Alpha Five desktop?

V8 will run on W7, although you will likely need to run in compatibility mode, with one major caveate - the alphaprinter driver in 8 won't work or work right in W7.

If there are very many users, moving to a dedicated server with a server os should be more stable and faster. But if you have to make a sale to get the server, get a good one - processor and drive rpm speed. I always try to sell 15,000 RPM SCSSI's.

a gigabit network is defintely faster then a 10/100 - especially if they also upgrade the cabling.

[DaveM]

Re: What happened to Alpha Five desktop?

so far, I have noticed NO difference in speed going from v7 to v10/v10.5 except for a faster initial load time in v10/v10.5. I have limited testing with 9 users(my beta sight) with peer to peer. Noticed no difference there either.

Maybe I am doing different things than you though.

This particular app has about 24 tables and numerous sets. Very limited indexes(usually no more than 4 per table). there are about 9000 records being worked with or stored.

.

[Jeff Fried]

Re: What happened to Alpha Five desktop?

V8 will run on W7, although you will likely need to run in compatibility mode, with one major caveate - the alphaprinter driver in 8 won't work or work right in W7.

If there are very many users, moving to a dedicated server with a server os should be more stable and faster. But if you have to make a sale to get the server, get a good one - processor and drive rpm speed. I always try to sell 15,000 RPM SCSSI's.

a gigabit network is defintely faster then a 10/100 - especially if they also upgrade the cabling.

There are maybe 4 users at a time, with two or three handling orders simultaneously. Hard to argue with a Windows server. The server is dedicated, but peer-to-peer (i.e., not a server OS) and has served them well for years. It's just that nagging difference in performance using V10 instead of V7 that creates this stop in going live with the new system.

Jeff

[JohnZaleski]

Re: What happened to Alpha Five desktop?

One thing to check.Make sure in Settings under network the network refresh is disabled in v10.
Also play around with disabling your antivirus maybe v10 is more affected by your virus software than is v7.

[Jeff Fried]

Re: What happened to Alpha Five desktop?

One thing to check.Make sure in Settings under network the network refresh is disabled in v10.
Also play around with disabling your antivirus maybe v10 is more affected by your virus software than is v7.

Regarding the virus software, I did disable it -- in fact, uninstalled it at one point -- to test its effect. No difference. I also just checked your suggestion regarding network refresh and it is set to 0 (disabled).

Thanks for the suggestions.

Jeff

[Jeff Fried]

Re: What happened to Alpha Five desktop?

I'm surprised you never went to V8 - which was appreciably faster than V7.

I have not done extensive micro testing (note that Ira Perlow has a utility that will help you measure times for scripts and parts of scripts for tracing where the differences may be), but I recently moved a big app from 8 to 10 and did not notice any difference in performance - running on Server 2003.

In preparing to move my biggest app from V8 to V10, I tested in both, where it rebuilds all the indexes from scratch running through about 360 tables, and the timing was the same, XPPro - but not faster.

I've had a chance to look at V8 a little closer and, so far, it behaves on the "old" application similar to, if not better than, the production environment that uses V7.

I tried looking at it running my new application, written with V10, and it may also be working normally, though with the caveat that it will need to be re-written to go backwards from V10 to V8. Things such as pop-up lists that should be coming up automatically are not doing so, so I need to push it for this testing. But it is certainly not exhibiting the stalls and slow behavior of V10. I just don't understand why V10 is behaving as well under V10 as V8 for your application and I'm not seeing success with V10.

If it weren't for the testing in which the only variable being tested is V10 vs. V7 using the existing production application, where V10 is noticeably slower, I would suspect my new application as being the issue, or the way I initiate the order entry screen, or pop-ups, or something like this. But the old application is so vanilla, and yet V10 is an issue. However, as I said, V8 is not an issue. And if, as you say, V8 was faster than V7 and will be compatible (to the extent you mentioned) with Windows 7 (it's 64 bit, by the way -- hope that still holds true), then the answer I'm looking for may be V8 for this client. I can't go back to V7 due to the W7 constraint.

I thank you for pointing out your experience with V8 and leading me to a potential solution.

Jeff

[Peter.Greulich]

Re: What happened to Alpha Five desktop?

I'm surprised you never went to V8 - which was appreciably faster than V7
....
from 8 to 10 and did not notice any difference in performance - running on Server 2003.

I have to concur 100% w. Martin. V8 is significantly faster than v7, and v10 is not slower than v8 - in my experience. So, unfortunately, I can't shed any light on your particular problem.

[perry9999]

Re: What happened to Alpha Five desktop?

I have your answer!!

I ran into the same thing with v10 vs v8..

The problem is the browse lists use way more overhead, and slow down a lot with many records..

HERE is your answer,

If you don't have an autoexec script, create one..
(called autoexec)

in the autoexec script, run this xbasic command:

a5.System_Mode_set("index_pseudo_position_cutoff","1000")



this disables the scroll bar feature that sucks up all the time.

Here is the response I received from the Bug report team:


There is a significant difference between V8 and later versions of
>A5 when browsing a table in index order. In V8, the vertical scroll
>bar does NOT reflect the relative position in the browse, and cannot
>be used to accurately position the cursor on a particular row in the
>browse. In V9 and V10, the vertical scroll bar does reflect the
>position in the browse and can be used to move to a particular place
>in the browse. Furthermore, when you move the vertical scroll bar,
>you get bubble help showing you the key values. This is very helpful
>in positioning the cursor in the browse.
>
>This behavior of the scroll bar, and the bubble help showing key
>values, does not come for free however. It does represent an
>additional processing burden. In most cases, this extra burden is
>not noticeable.
>Apparently on your system (but not mine), it is extremely noticeable.
>
>Therefore you might want to test turning OFF the advanced scrollbar
>features in the browse.




*********************************************************
[QUOTE=Jeff Fried;540599]I am so dismayed to find out that there is a vast difference in performance between V10 and V7 (yes, I'm skipping V8 and V9 for this analysis because V7 is what has been running in the live environment).

The project was to rewrite an order entry application that now includes many bells and whistles. At the same time, computers were upgraded along with the server. The computers started with Vista 64 bit but were upgraded to Windows 7 64 bit. But factor all that out and just focus on the variable V7 vs. V10. In a type of "laboratory" environment I set up with several of my client's computers so I could take my time to investigate why the revised app is not ready for prime time, I studied many combinations.

But in the end, using XP with a vanilla server, I discovered that my client's old application -- the one they currently use in production, and have been using for many years -- running it under A5V10 has a noticeable and substantial slower performance and response than running it under A5V7. I did many double-takes on this. At its basic level, just starting a new order which pulls up a list of customers to choose from, then proceeds with a pulldown of products to select from, is substantially different between V7 and V10. In V7, all is immediate response. In V10, hesitation. Not a big hesitation, but one which, when four or so people are pounding orders in, there will be a major performance hit to their operation. As to the revised application, loaded with many more capabilities, I expect disaster. Unfortunately, I can't yet test the new app in V7 vs V10 since it won't run in V7 (designed in V9, then V10, so refuses to run under V7). I'm thinking of rewriting the main data entry screen using V7 just to be able to test the new app under V7 vs. V10.

I doubt anyone has a solution to this as I have boiled this down to barebones testing. There is definitely, without question, a major decline in Alpha Five desktop from V7 to V10. I've had it with the V10 hype! I pass along the promise of the newer version to my clients, then end up with an unworkable solution in the end. Has Alpha Software's focus on wep apps killed Alpha Five desktop?

-Perry William