Registered: Sep 1999
Location: at work...
Posts: 736

I thought it would be interesting to start a poll of performance benchmark results on Visor Prism vs. Visor Platinum (vs. Visor Deluxe).

Here are my results, using both the QuickBits index (2bitsoftware.com) and Benchmark 2.0 (quartus.net):

Visor Deluxe:
QuickBits Index: 107 (without any speedups/hacks)
QuickBits Index: 324 (with QuickBits speedups only)
QuickBits Index: 655 (with QuickBits speedups & Afterburner)
Benchmark 2.0: 98% (without any speedups/hacks)
Benchmark 2.0: 124%(with QuickBits speedups only)
Benchmark 2.0: 260%(with QuickBits speedups & Afterburner)

Visor Prism:
QuickBits Index: 76 (without any speedups/hacks)
QuickBits Index: 292 (with QuickBits speedups only)
QuickBits Index: 295 (with QuickBits speedups & Afterburner)
Benchmark 2.0: 124%(without any speedups/hacks)
Benchmark 2.0: 148%(with QuickBits speedups only)
Benchmark 2.0: 241%(with QuickBits speedups & Afterburner)

(Note that since the Visor Prism uses the new 33MHz StrongArm Processor, I'm not sure how much of a difference afterburner makes -- especially since the max speed is only '32MHz' anyway -- although it definitely seems to affect the Benchmark 2.0 numbers)

According to the QuickBits index, a normal Visor Deluxe should rate at 107, which mine does. However, according to Benchmark 2.0, the Visor Deluxe *should* be 145%...I wonder why mine was so low without the speedups/hacks? Also, the Benchmark 2.0 indicates the Visor Platinum is rated at 204%, but I couldn't test it as I don't have one. =)

As for the most important indicator, actual perceived performance, I would have to say my Prism is not quite as fast as my Visor Deluxe. Maybe 16-bit color requires a lot of processing overhead...

[Edited by Rob on 10-19-2000 at 03:06 AM]

Registered: Oct 2000
Location:
Posts: 10

Here's what I got from my Visor Platinum:

QuickBits Index: 136 (without any QuickBits speedups)
QuickBits Index: 471 (with QuickBits speedups)
Benchmark 2.0: 204% (without QuickBits speedups)
Benchmark 2.0: 243%(with QuickBits speedups)

Registered: Oct 2000
Location:
Posts: 31

quote:

---

Originally posted by Rob

(Note that since the Visor Prism uses the new 33MHz StrongArm Processor, I'm not sure how much of a difference afterburner makes -- especially since the max speed is only '32MHz' anyway -- although it definitely seems to affect the Benchmark 2.0 numbers)

---

Registered: Feb 2000
Location: Moved to Clie Land
Posts: 331

Thanks KRYPTON, I should be getting my Platnium today and those numbers sure look good!

Registered: Feb 2000
Location: SW Virginia
Posts: 290

quote:

---

Originally posted by Rob
Visor Deluxe:
Benchmark 2.0: 98% (without any speedups/hacks)

---

Registered: Oct 1999
Location: Vancouver, BC, Canada
Posts: 36

I agree with potter, with my Visor Deluxe, and running afterburner 3 at 28MHz, my visor clock in at 275-278%. When using more than 1 overclockers (altho i guess quickbits technically isnt' one), I have in the past expereinced "loss" of increase.

I would suggest that doing a soft-reset first and then installing the quickbit and afterburner to see if they truly do "interfere". I have a feeling the order of activating them might also be important.

I am usually very fast to jump on any new "overclocking" programs but after a quick read of the users' comments on palmgear:

http://www.palmgear.com/software/sh...amp;prodID=9556

I think I will wait a couple days... just can't afford to hard crash these days...

[Edited by Gretz99 on 10-19-2000 at 10:52 AM]

Registered: Sep 1999
Location: at work...
Posts: 736

quote:

---

Originally posted by potter

quote:

---

Originally posted by Rob
Visor Deluxe:
Benchmark 2.0: 98% (without any speedups/hacks)

---

Your benchmarking of the Visor Deluxe is not accurate. What I think happened is that several of the over-clockers, when disabled, return the Visor to the normal state of for a Palm IIIx. The Visor has a few speed ups above that. I understand that after disabling, a soft reset will return the Visor to it normal state. I do not use any speed ups, and running Benchmark 2.0 I get 145%, as is displayed earlier in the graph.

---

Registered: Oct 2000
Location:
Posts: 6

These are MY benchmarks in ePocrates:

I searched for Migraine under indcations:

34 seconds Deluxe : 21 seconds Platinum

Search for Dupont under manufacturer:

19 seconds Deluxe : 12 seconds Platinum

Search for Rhabdomyolysis under adverse effects:

149 seconds Deluxe : 84 seconds Platinum

38% improvement

37% improvement

44% improvement

I know it's not scientific, but I love the Platinum.

Michael

Registered: Feb 2000
Location: Moved to Clie Land
Posts: 331

Nothing has blown up yet....

Ok, I have the Platinum and have 20% memory available, just installed afterburner(set at 32) and Quickbits. The Quickbits index is at 464
and Benchmark is at 235%

So why are my numbers slower than what you ae achieving? I realize that different units running different Apps could be the culprit, but I'm a long way off from 655. I plan on playing with the settings some, and esp try the Stowaway and other possible problem areas, but my primary purpose is to get the best possible performance out of my unit while still maintaining a safe level for longevity. I have no qualms with paying the fees, but I want to get this fine tuned before I fork over the cash...

Any suggestions would be greatly appreciated.

__________________
"Stupid Handspring."

Registered: Jun 2000
Location: Sunnyvale, CA
Posts: 216

Visor Deluxe vs Prism/Platinum

After reading James' review of the Prism and seeing his benchmark numbers, I noticed that, while faster than the Deluxe, the Prism does not seem as fast as its 33MHz clock speed would imply.

Here are my benchmark numbers for a standard Visor Deluxe:

Benchmark 1.5: 158% at standard 16MHz with no Afterburner and no Quickbits

(NOTE: The new Benchmark 2.0 version seems to give slightly lower numbers.)

Pocketmark 0.2.1: at 16MHz, no Afterburner, no Quickbits
IntMarks: 128
MemMarks: 101
FloatMarks: 131
PlamMarks: 124 <- the overall score
DBMarks: 49 <- optional test
GraphMarks: 112 <- optional test

Now when I use Afterburner 3.0q at 28MHz, no Quickbits I get:

Benchmark 1.5: 273%

PocketMark 0.2.1:
IntMarks: 220
MemMarks: 175
FloatMarks: 228
PlamMarks: 214 <- the overall score
DBMarks: 84 <- optional test
GraphMarks: 193 <- optional test

This is an average of a 173% speed increase by going to 28MHz. Note that this is FASTER than the Prism (and I assume the Platinum) with a PocketMark overal score of 192 (from James' review). This does not seem to make sense as the Prism's processor is running 5MHz faster (18%) than my overclocked Deluxe. Yet it is 10% slower! Could there by some effect from Palm OS 3.5.2?

Here's another idea: The Deluxe normally runs the screen at 1-bit depth. The Prism runs in color which is at least 8-bit depth. But the benchmarks, with the execption of the GraphMarks test of PocketMark, do not draw anything. So I'm not sure if this would have much of an effect on the scores. More testing is called for.

NOTE: I do not yet have Quickbits to test with.

I would love to see some comparisons of the PocketMark GraphMark score of the Deluxe straight, overclocked, with Quickbits, and overclocked with Quickbits, compared with the Prism both straight and with Quickbits. The Prism tests should also be run at both 8-bit color depth and 16-bit color depth.

Can someone also run PocketMark on the Platinum. I haven't seen this data posted anywhere here yet.

Registered: Dec 1999
Location:
Posts: 97

Why does my VDX get slower results on the QuickBits Index when I turn on Afterburner? Are these results reliable?

VDX with QuickBits & Afterburner off: 120
VDX with QuickBits on & Afterburner off: 324
VDX with QuickBits off & Afterburner on: 99 or 112*
VDX with QuickBits & Afterburner on: 301

*The variable results are from different tests; I've gotten both results numerous times.

I'm using Afterburner v3 at 28MHz with these settings (and, yes, I perform a soft reset after disabling Aftterburner):

0-ws Core: On
Fast Mpx: On (same speed when Off)
Fast Page: On
Slow RAM: Off
Fast Hotsync: On (same speed when Off)
Adjust ticks: On (same speed when Off)
Default: Norm
Find: Norm
LCD/n: 1
Notify on: Off
Gray: Off
Accel: Always
Slow: <20%
LClk/2: 24MHz

Final question: What I'd most like to speed up is my Backup module. I get a very modest improvement from QuickBits (from 5 minutes to 4-and-a-half), but none with Afterburner. Any suggestions to speed it up? Will Platinum improve backup speed significantly?

Registered: Sep 2000
Location:
Posts: 39

i dunno. but heres what im gettin. (i dont use afterburner, cause theres no need.)

qb off: qbi: 136
benchmark: 204

qb on: qbi: 468
benchmark: 244

im loving quickbits! now ive got the fastest palm pda out there. i dont use afterburner, because even though in tests it may appear to make it faster im just dandy with my 33 mhz DragonBall HS, or whatever its called.

Registered: Sep 2000
Location:
Posts: 39

i just ran palmmarks on my VISOR PLATINUM.

without QuickBits:

intMarks: 202
MemMarks: 178
FloatMarks: 201
PalmMarks: 198
DBMarks: 346
GraphMarks: 107

with QuickBits:

intMarks: 229
MemMarks: 451
FloatMarks: 231
PalmMarks: 263
DBMarks: 364
GraphMarks: 113

Registered: Oct 1999
Location: Vancouver, BC, Canada
Posts: 36

Just an observation regarding the significant increase in the DB marks component of the Visor Platinum:

bpowers's Platinum clock in with a DB marks in the 300s, while my VDx that is Afterburnt at 32 and Quickbitten only has a DB mark of 67. As well, MPM's VDx's DB mark is also < 100 in his benchmarks even with Afterburner.

Other than that, my VDX clock in at the following:

Afterburner 3.0q at 32, Quickbits All options on:

IntMarks: 258
MemMarks: 543
FloatMarks: 264
PalmMarks: 302
DbMarks: 67
GraphMarks: 224

Benchmark 2.0 at 257%

Quick Bit Index at: 655

It would appear most of my benchmarks are pretty much on par with the maximum that a VDx can achieve, with the exception of the DB marks component, which seem to generate more fluctuation between units- any body has an idea on this?

I am not sure if my interpretation is correct but as it stands right now, it would appear to me that PalmMarks and QBI wise, the Platinum still is behind the VDx (due to the lack of a version of Afterburner that can take advantage of the Platinum's processor speed).

Having said that, I hope I will be proven wrong soon with a new/improved version of Afterburner.

Registered: Jun 2000
Location: Sunnyvale, CA
Posts: 216

Wow. I'm surprised by two things:

1) The DBMarks from the Platinum are very large. So much so that I think there must be something we are not understanding about the test. Could it be sensitive to how much RAM is being used? My Visor Deluxe had 4529K used - over half full. Could it be the new Palm OS 3.5.2?

2) The speedup in the GraphMarks test on the Platinum is disapointing. I was expecting Quickbits to speed up the graphics much more than that.

One thing I did notice is that MemMarks number from the Platinum is not significantly higher than an overclocked Deluxe - 178 on the Platinum and 175 on my 28MHz Deluxe. This gives me some assurance that Afterburner is not speeding up the memory accesses beyond the maximum speed of the memory chips. This was my primary worry about using Afterburner.

Still it seems that the Visor Deluxe, when overclocked to 28MHz and with Quickbits added, is faster than the Platinum.

Personally I'm dissapointed in the Platinum. No case color options and it's really not much faster then the deluxe. I don't think that just Palm OS 3.5.2 is worth the extra $50 this unit costs over the Deluxe. If the speedup had been closer to 100% like the 33MHz clock rate implies...

Registered: Feb 2001
Location:
Posts: 91

I don't know how but

I got 306%!!!!!!!!!!!!! More info to come....

__________________
RIG | HEAT | MAIL
<br><div align="center"><EMBED src="http://xzaver.tripod.com/Xzaver/X.swf" WIDTH=500 HEIGHT=30></div><BR></BR>
ClieSource.com You guide to everything Cli�

Registered: Apr 2001
Location: Chicago, USA, Earth
Posts: 11

Here are my Prism results. I am using AfterBurner 3.0u, I have Fast Page checked off, 0-ws Core does not affect my score when I have the clock pegged. I am able to achieve stability at up to 43MHz.

Anyway, here are the results.

PocketMark 0.2.1

IntMarks 304
MemMarks 171
FloatMarks 301
DB Marks 254
Graph Marks 63

Palm Marks 284

****

Benchmark 2.0 325%