Erm is it me or did the A1G3SE at 600Mhz trounce
the Pegasos 1 at 600Mhz?

Or am I reading it wrong?

Whew!

@ DaveW

Quote:

Erm is it me or did the A1G3SE at 600Mhz trounce

These tests clearly suggest that the A1G3SE outperforms the Pegasos while using a G3 at the same clockrate.

The testers are planning to do more tests however and are open to suggestions.

First thing to determine would be the front side bus speed settings.

I surmise the A1 is running at 133 Mhz FSB but what about the Pegasos?

You'd expect to see the same performance really.

OTOH, claiming that a G3/600 could outperform a G4/800 was always a complete joke.

_________________

Well this is news worth celebrating. We actually
have some data that puts the whole debate on a firm footing ( apart from the jam tomorrow promises
that muddy the water ).

Interesting reading.

_________________

About the G4-Xe, what kind of an speed increase can one expect with Altivec optimisation?

The Pegasos benchmarks have now been updated, the differences between between the PegasosG3 and AmigaOne G3 results seems to have been mainly a L2 cache issue. Ole-Egil suggested this to Gunne and apparently has been able resolve the issue.

This is exactly what I stated when BB shared his wild claims with the public. The PegasosG3 and AmigaOneG3 @ 600 MHz are about equivalent with regard to Linux performance.

@ Eric_Z

The performance gains greatly differ for different types of applications. AltiVec adds 128-bit-wide vector execution units to the PowerPC architecture. Here you can find a well done introduction.

Makes sense really I suppose. Good to see that testing is underway. Makes it all seem real I suppose

_________________
Folding@Home Team AmigaWorld no: 33424

My word exactly.

it's one thing to know that the boards are in the hands of beta testers (and others) but seeing actual benchmarks from the machines. is Neat, one picture (or benchmark page) is worth more than a 1000 words.

btw: my first post on this page, nice to see the more civilized nature of people here.

_________________

Perhaps someone could do an explanation of those tests for the less techie amongst us?

Any takers?

_________________
Folding@Home Team AmigaWorld no: 33424

I find it rather funny that, although the Pro-mos group are quick to jump the gun and dismiss these benchmarks as useless. None of them have posted this "news" item on any Morphos news site. (yes i know it's not morphos test, but Pegasos don't have a seperate news site).

(Disclaimer: not a troll attempt, or flame bait)

_________________

Just done a quick work per MHz comparison and on average the PPC's do more work then the Athlon. OK so the Athlon still beats the PPC as it is running faster however, if the PPC ever reaches the speed of the Athlon/P4 it should at least give it a run for its money

(PPC on average gave 13.5% more punch per MHz)

Surprisingly tho the G4 performs worse per MHz in 50% of the tests compared to the Athlon and 90% of test when compared to the G3!

_________________

Quote:

Surprisingly tho the G4 performs worse per MHz in 50% of the tests compared to the Athlon and 90% of test when compared to the G3!

I could be wrong but I seem to recall a thread where it was explained that the G4 relies heavily on the AltiVec part for it's speed and so, if this is not in use, it's performance is somewhat crippled.

I'm not a PPC guru so I couldn't claim that as fact - just how it's been explained to me.

Perhaps someone can confirm/correct?

_________________
Folding@Home Team AmigaWorld no: 33424

Quote:

Just done a quick work per MHz comparison and on average the PPC's do more work then the Athlon. OK so the Athlon still beats the PPC as it is running faster however,

Note that "Athlon XP" also has undergo it?s own IPC improvements i.e.
1. Barton core; 333DDR FSB + 512Kb L2.
2. last Athlon XP core change; 400DDR FSB.
3. nForce 2 chipset. i.e. getting away from the relatively crippled VIA KT400 ?crapset?.

This is not factoring Athlon64 IPC improvements over Athlon XP; i.e. x86-64 instruction set, SSE-II, significantly improved branch prediction, a larger L2 cache and a slightly longer pipeline to allow additional clock ramping (12 stages rather than 10 in the Athlon).

To illustrate my point, refer to
http://www.amdboard.com/hn03130301.html
http://www.amdboard.com/barton3200.html

The documents revealed the following;
0. AThlon 64 @2.1Ghz is rated at '3800+'.
1. Athlon 64 @1.8Ghz is rated at '3200+'.
2. Athlon XP @2.167Ghz(Barton) is rated at '3200+'.

In trems of IPC, AThlon 64 > AThlon XP(Barton).

Athlon?s Rating is based on the equivalent speed of the older Athlon Thunderbird core. Which indirectly compares with late model Pentium CPUs.

IPC alone is quite pointless without factoring the whole package.

Quote:

if the PPC ever reaches the speed of the Athlon/P4 it should at least give it a run for its money

Are you assuming that AMD will forget it?s own IPC improvements?

For Intel's IPC improvments refer to
http://www.inquirerinside.com/default.aspx?article=8519

Quote:

The problem isn't with the processor exactly, it's a real firecracker from all accounts. But the Intel marketing droids have spent years trying to convince everyone that megahurts matters. Now an Intel processor has come along that proves the marketing droids were wrong. The 1.6GHz Pentium M is performing more like a 2.4GHz Pentium 4. Now adverts are starting to appear for machines using the new processor but the clockspeed is conspicuous by its absence, as you can see over here. It seems that the Intel marketing droids have tied their own laces in a knot. Will the company be launching a PR rating of its own? Will it finally admit that the megahurts wars didn't make any sense? Surely it can't keep the Pentium M speed out of adverts forever, what happens when the next model comes out? What is the official line when a customer asks how fast it is? Grab yourself some popcorn because this should be fun. µ

In
http://www.inquirerinside.com/default.aspx?article=8386

Quote:

The Banias (Pentium M), he added, is definitely based on the Pentium III. The principles used are the same as the Pentium 4 moving to the Prescott architecture, with Intel making a few microarchitectural changes to improve performance and adding performance tweaks to make the Pentium III a little bit faster.

PS; I?m well aware of PowerPC 970.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)