Personally I would go with a G4 MDD for simplicities sake, as a Dual 1.25 or 1.42 G4 will provide plenty of horsepower for what you want, and you tend to get a much nicer graphics card than most of the Sawtooths included (My second one came with its original Rage 128 with 16MB VRAM, which although great in '99 isnt so great anymore), and they can run 10.5 with ease. (I currently have a Dual 867 as my primary PowerPC machine, and like the G4s so much Im planning on getting either a second MDD or third Sawtooth/Gigabit to be my file server), and the MDDs have the advantages of being Dual Optical Bay capable and have room for 4/5HDDs and an Optical Drive without having to add adapter cables to hook up the drives (These are all good features to have down the road when you fill it up with stuff).
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Help with choosing the right G4?
- Thread starter Miguey
- Start date
- Sort by reaction score
See, I'm just saying facts. I don't have to say for how long I've been using G4s nor anything, because that would not make facts become "more true".
I don't seem to have any real world experience? Well, that ain't true. But in my opinion, its the OP's job to find out what to do. Saying too much of personnal experience is NOT objective.
If you like so much real world tests, why would you try to prove something using benchmarks? (aka "thing far away from real world experience").
If you want me to talk about personnal experience, here's one thing I've found out. Xbench is really bad for testing the GPU. On my current mac desktop upgrade project, it gave the exact precise same score with a stock radeon 9000 pro and my newest 7800 GS overclocked, using latest Xbench version. However, a video test that tries to play video as fast as possible without lagging showed +100% better performance, which is hell of a lot. Playing a real video, how closer can we get from real world testing?.... So do not put too much trust in Xbench.
Xbench is best at testing the result of CPU + hard drive performance.
4X agp is built to be faster than 2X agp, that is FACT. Unless your video card is too slow to use the bandwith of 2X agp, then that would not make any difference.
You seem to get mad a little too fast. I mean, this is just a forum for helping some dude. This is not the place for personnal attacks.... Stay cool man...
You can't sit and use my G4 so all I can offer you is benchmarks.
How many G4 towers have you owned? How long have you been using them in total for? How many of the 6 different models have you actually owned and used? What other than basic specs do you have to offer?
Regardless of how good or bad xbench is at GPU scoring it's isn't going to make AGP2x the same speed or faster if it's in fact slower. Most of these AGP cards are actually intended for AGP 8x so in any G4 they would under perform. My Sawtooth with G4 1.8ghz single and Geforce 6200 games as well as my former dual 1.42 MDD that had a Radeon 9600.
You base everything you say on specs and therefore have no real argument to offer when it comes to G4 towers. I'm also certainly not "mad". I simply don't like people that try to act like they know what they are saying. It's misleading advice when you just state specs.
How many G4 towers have you owned? How long have you been using them in total for? How many of the 6 different models have you actually owned and used? What other than basic specs do you have to offer?
Regardless of how good or bad xbench is at GPU scoring it's isn't going to make AGP2x the same speed or faster if it's in fact slower. Most of these AGP cards are actually intended for AGP 8x so in any G4 they would under perform. My Sawtooth with G4 1.8ghz single and Geforce 6200 games as well as my former dual 1.42 MDD that had a Radeon 9600.
You base everything you say on specs and therefore have no real argument to offer when it comes to G4 towers. I'm also certainly not "mad". I simply don't like people that try to act like they know what they are saying. It's misleading advice when you just state specs.
If it gives the same EXACT SCORE for a graphics card that is 4 times better, why would you think its good at rating AGP 2X vs AGP 4X? That is non-sense.
It looks like Xbench just finds the model in its database and throws in the rating of the stock AGP card. That would not be the first bench to do this.
If you don't like people that try to act like they know what they are saying, I don't either. Who likes that anyways?
On my side, I don't like when people boast about their experience and stuff... The reason why you seem to think I know nothing is because as I write, I expect the reader to be bright enough to interpret the facts, using his own knowledge. In other words, what pisses me off is people that talk to others like if they were total noobs, so I'm not doing it with you.
It's up to you to think I'm stupid, or trust the infos and investigate. It would be useful to get a more reliable bench for AGP graphics. It is also up to you to test different AGP cards and see what happens. That way we could see where the problem is. Maybe Xbench just gets the stock info for GPU, maybe theres an issue with some cards... , maybe it's just with the 8X cards, or the flashed pc cards, or ... etc.
edit: as I said, 2X AGP vs 4X AGP will NOT make a difference if the card the OP is using doesn't come close to saturating the bus. However, a 7800 GS like mine will, and as people have reported, it can even come close to saturate an 8X AGP bus. Google it if you don't trust.
--->>>So the advice would be... if you plan to play Doom 3 or Quake 4 at decent quality and FPS, definitely stay away from 2X AGP. If not, I guess it doesn't really matter.
Last edited:
Having used a 1.25GHz SP MDD, a 450MHz DP Gig-E, and a 800MHz DP QS with a number of different video cards, the 2x/4x AGP slot doesn't make a practical difference at all for the video cards you're going to be using, unless, by chance, you're going to use a 6800 Ultra DDL with two dual-link displays. However, I don't think you'd be able to power a 6800 Ultra DDL with the stock power supplies of any of the 4x machines. I know that an OC'd FX5200 Ultra significantly drew on the power supply of my QS. Hell, I doubt that big-ass card would even fit in a G4 case.
OP, if you're willing to mod, a Sawtooth or GE is going to be your best bet. Otherwise go for the MDD. The MDD will be your best out-of-the-box experience.
OP, if you're willing to mod, a Sawtooth or GE is going to be your best bet. Otherwise go for the MDD. The MDD will be your best out-of-the-box experience.
Nova:
Download this benchmark and run it on your MDD. It's the absolute best GPU benchmark on Mac.
I will run the same on my Sawtooth and post a screen shot of the results here. I have ran this before but not for a long time.
Edit:
I ran it twice on my G4 at different res.
@1680x1050
@1920x1200
My MacBook with integrated GMA 950 GPU only scores around 900-1000.
Download this benchmark and run it on your MDD. It's the absolute best GPU benchmark on Mac.
I will run the same on my Sawtooth and post a screen shot of the results here. I have ran this before but not for a long time.
Edit:
I ran it twice on my G4 at different res.
@1680x1050
@1920x1200
My MacBook with integrated GMA 950 GPU only scores around 900-1000.
Last edited:
Mine is a single 1.8. Not a dual you can see that in the Xbench system specs. If would be a great way to tell if your AGP is faster as all it tests is the GPU and graphics throughput.
@Nova and Zen-State
Before this goes any further, gaming wise, the most I'll be installing on it would be Warcraft 3 and Starcraft Broodwar. So graphics aren't exactly a priority. I do appreciate your guys' explanation though! I'm pretty sure if I was doing some hardcore editing it would take place but for now I'll suffice with the stock graphics if it came with it.
@SuperJudge
At this point I'm looking towards the MDD because it just needs me to last the next 2-3 years. eMac's are making their way up my list though because of the price(60 bucks, spend another 30 on memory upgrade and get an external) but the MDD is still winning.
Thanks again guys! I'm on the lookout every day for that one deal.
Before this goes any further, gaming wise, the most I'll be installing on it would be Warcraft 3 and Starcraft Broodwar. So graphics aren't exactly a priority. I do appreciate your guys' explanation though! I'm pretty sure if I was doing some hardcore editing it would take place but for now I'll suffice with the stock graphics if it came with it.
@SuperJudge
At this point I'm looking towards the MDD because it just needs me to last the next 2-3 years. eMac's are making their way up my list though because of the price(60 bucks, spend another 30 on memory upgrade and get an external) but the MDD is still winning.
Thanks again guys! I'm on the lookout every day for that one deal.
EDIT: Just so you don't have to read my long-winded post () Miguey, I'll just repeat here that I vote for the eMac or iMac. They should run all the software you've mentioned quite well, and are cheaper and more hassle-free than an upgraded G4 tower.
I think the key point is that, with an upgraded CPU, the only difference in performance is going to come from the bus and AGP slot. You can slap the same CPU in any of these machines, and (generally) the same video card. The slower bus on the Sawtooth is a disadvantage but a small one. The Yikes! G4 is the only one that is too "crippled" by it's hardware to be worth upgrading.
When upgrading my G4 Digital Audio, I swapped out my dual 533s for a single 1.4Ghz CPU, and installed a 256MB GeForce 6800GT (flashed). The video card never even came close to being stressed; the CPU and bus were by far the major limiting factor.
Getting back to the OP's question though, the eMac and iMac can probably get the job done much cheaper and quieter than an upgraded G4 tower, even though the latter will be considerably faster in some tasks. The GeForceFX 5200 in the iMac is adequate for just about any software you'd run on a G4.
I continue to be a big fan of the PowerMac G4s, they are great machines. But in this case I would vote for the eMac or iMac models the OP mentioned. They have a good CPU, adequate video performance and require no upgrades beyond memory.
) Miguey, I'll just repeat here that I vote for the eMac or iMac. They should run all the software you've mentioned quite well, and are cheaper and more hassle-free than an upgraded G4 tower.I think the key point is that, with an upgraded CPU, the only difference in performance is going to come from the bus and AGP slot. You can slap the same CPU in any of these machines, and (generally) the same video card. The slower bus on the Sawtooth is a disadvantage but a small one. The Yikes! G4 is the only one that is too "crippled" by it's hardware to be worth upgrading.
When upgrading my G4 Digital Audio, I swapped out my dual 533s for a single 1.4Ghz CPU, and installed a 256MB GeForce 6800GT (flashed). The video card never even came close to being stressed; the CPU and bus were by far the major limiting factor.
Getting back to the OP's question though, the eMac and iMac can probably get the job done much cheaper and quieter than an upgraded G4 tower, even though the latter will be considerably faster in some tasks. The GeForceFX 5200 in the iMac is adequate for just about any software you'd run on a G4.
I continue to be a big fan of the PowerMac G4s, they are great machines. But in this case I would vote for the eMac or iMac models the OP mentioned. They have a good CPU, adequate video performance and require no upgrades beyond memory.
Lord Blackadder:
True about the Yikes. It has a G3 logic board and a much slower memory controller. Even though it has the same 100MHz bus speed as the Sawtooth and GE it's much slower.
I personally have never liked all in one computers like eMac's and iMac's. If the screen goes you're pretty much stuck with a Mac brick. A tower is much easier to replace parts in, far more expandable and also doesn't rely on a built in screen so you can you exactly the screen you want.
True about the Yikes. It has a G3 logic board and a much slower memory controller. Even though it has the same 100MHz bus speed as the Sawtooth and GE it's much slower.
I personally have never liked all in one computers like eMac's and iMac's. If the screen goes you're pretty much stuck with a Mac brick. A tower is much easier to replace parts in, far more expandable and also doesn't rely on a built in screen so you can you exactly the screen you want.
Aquamac was able to fit in an Nvidia 7800 GT 512 MB in a G4 MDD. Google it
That 7800 GT isn't quite as big as the 6800 Ultra DDL AFAICT, but that's a pretty impressive bit of kit and I love that it's got a single slot cooler. For those too lazy to Google, look here.
This is precisely what I hate about all-in-ones and why I generally would hesitate to recommend them to anyone.
The high voltage board in my 266Mhz iMac blew once, and the collateral damage fragged the power supply and the CRT. This was in the days before the Mac Mini, and my only option was a $400 repair. After that I swore I'd never buy an all-in-one again.
However, LCD/LED screens are smaller, lighter, longer-lived generally more reliable than CRTs, so I'm somewhat back on the all-in-one bandwagon. G4 towers are bigger and much louder, and upgrading them can be quite expensive. The only original internals in my G4 tower are the logic board and PSU, everything else was ugraded...that's a lot of new hardware to buy.
It all depends what you want. Yes, my G4 can handle dual 27" displays and I've got 4 HDDs in the thing, but there is something to be said for the lower cost and lower footprint of the AIOs. For some users they are the best option.
Why didn't you at least show your screen snap? When you run it again make sure you capture the windows when it's done. Just typing the results doesn't count.
@1280x1024
LOL yeah, I know its stupid. I must say its kinda messy out there, having my MDD temporarily at my parents house and using my laptop at home 1 hour away. I had the chance to get to my MDD and run the test, but I was too much in a hurry so I didn't post my screenshot yet. I can tell you the score is 14450. Believe it or not, you will see the screen tomorrow, when I get back to my parents house. I must say I was surprised to score that high... and atm I have a hard time to believe it, but anyways... Check it out tomorrow
Thanks for posting score @1280x1024, that way the bench comparison means something.
Edit: Ok maybe im too tired to remember the exact score, but I do remember counting the number of digits, and it had 1 more than yours, and thats what surprised me.
The 7800 GT is a dual slot card. Aquamac modded it by changing the heatsink. So yeah thats a pretty big card since it is supposed to take 2 slots.
----------
This is the results for my PowerBook G4 1.67 ghz, 1 GB RAM.
This computer actually feels like a stock MDD (with a little more RAM), and extensive use showed me the GPU in this laptop is almost (in terms of performance) the same as the Radeon 9000 Pro 64 MB (stock in G4 MDD).
The card is an ATI Radeon 9700 Mobility 64 MB (bench shows the wrong model, idk why.)
MacTracker shows the powerbook g4 I have uses 4X AGP interface.
It scores almost as high as your Sawtooth, with 2X AGP and much better video card.
Do the math...
I'll post my MDD's results later tonight.
Are you saying a Geforce 6200 is a "much better video card" than the Radeon 9700? It isn't. A 9700 would even be faster than a GeForce 6600. Even if it is the mobility GPU. The later PowerBook G4's had very good graphic capabilities and you can see in your test.
I can also tell you're running Tiger from that window border vs. my G4 running Leopard. Back when I ran Tiger on my G4 with the same CPU/GPU I got over 6000. Not really a fair comparison with a different OS.
I ran it on my MacBook to show how little a role the CPU really plays in the score. It's all GPU performance. My MB would smoke any of the G4's we're talking about here CPU-wise but as you can see in this test that the Intel GMA 950 is pretty gutless when it scores 1/5 what a G4 does.
MacBook C2D 2 GHz running 10.6.8:
Why don't we take this discussion to the thread I made last night here for OpenMark. You can see it near the top and it's titled "OpenMark results on PowerPC Macs". I don't want to keep dominating this guys thread with benchmarks.
Last edited:
+ it has gigabit ethernet over the 10/100mbit/s AGP G4
(yes I know you can upgrade it with a PCIcard, but would I want to do this, when all slots will be closed and sharing the bandwidth perhaps the impact would be not measurable, but its nice to have it already on board, huh?)does anybody clearly know whether the 9800 is OS 9 compatible? I read people write the RV200 chip was the last ATI (R7500/8500/9200) to support OS 9 and themacelite does list it as unsupported, too. But I saw descriptions on Ebay saying the 9800 works fine under OS 9. Others then again say it kernel panics...
I found a Dual 1,8GHz (not the MDD version) on Ebay going for 126,-EUR, but I was to hestitant with bidding...
you can use the external output to run a monitor with some iMacs (and perhaps eMacs, I did not care) and there are nice Mods taking the CRT out and replacing it with a LED or you could mod yourself a case and put the iMac/eMac inerts in it and connect the external-VGA-out to a monitor.
But you are right expandability wise.On the CPU upgrade, does it actually make sense to up an MDD 1,25GHz/1,42GHz to 1,6 or 1,8GHz? (noticable) I think an upgrade for a former G4 from lets say 500Mhz to 1,6GHz would make a better bargain (if you are not doing it as your hobby and are willing to spend much money anyway).
Last edited:
The Radeon 9800 Pro will function in OS9, but with no hardware acceleration. If you want to have your cake and eat it too, you can put a Radeon 9800 in the AGP slot to use with OSX, and then stick a Radeon 9200 PCI video card in one of your PCI slots to give you hardware acceleration in OS9.
It depends. There are several different types of G4 CPU out there (7455, 7457, 7447, 7448), so direct clock speed comparisons are not always useful. Some (7455/57) have a big fat L3 cache, so the 7455 1.4Ghz CPUs may equal or exceed the performance of the 7447-7448 1.8-2.0Ghz CPUs in several tasks. There is also the issue of single vs dual CPUs. The dual CPUs don't help much with gaming but can offer a nice boost with some professional applications like Adobe Creative Suite.
Short answer? If you have one of the dual 1.25 or (especially) 1.42 machines, I wouldn't bother. Also, just for perspective, notice from these benchmarks just how much slower CPU performance is with the fastest G4s versus even the early Intel Macs.
I've done extended reaserches on GPU cards an mac os 9. It should work without any problems, but you won't get any hardware acceleration under OS 9. So forget about OpenGL games under mac os 9.
Edit: lol was too slow, somebody answered before me. I will add that the last and best AGP graphics card to be fully supported under OS 9 is the Geforce4 Ti 4600 128 MB.
----------
I second all what you are saying except that part...
Since OS X uses both CPUs at the OS level, it will make a difference to have dual CPUs. However some games have better optimization than others for dual G4 cpus.
Under OS 9, only 1 cpu will be used most of the time, unless you play Quake 3 arena, which back in the day was one of the rare games well optimized for dual G4s.
One further tidbit for you, Cox Orange - I believe that the rare GeForce4Ti (4600) OEM Mac edition video card will function with full hardware acceleration under OS 9. It is roughly equivalent to the Radeon 9800 in power, but was never offered as a retail card and is therefore rare and expensive (though flashed PC examples are cheaper). It is probably the fastest OS 9-compatible card ever made.
It does not, however, support Core Image in OS X, so if you plan on running both OSs your best bet is still a combo of the Radeon 9800/Radeon 9200 PCI cards.
Again, it depends. If you check out various benchmarking tests at Barefeats, you see that you get a small boost in games with the dual CPU machines, but the difference is often fairly small. It helps, no doubt, but it's not a huge difference.
It does not, however, support Core Image in OS X, so if you plan on running both OSs your best bet is still a combo of the Radeon 9800/Radeon 9200 PCI cards.
Nova77 said:
Again, it depends. If you check out various benchmarking tests at Barefeats, you see that you get a small boost in games with the dual CPU machines, but the difference is often fairly small. It helps, no doubt, but it's not a huge difference.
Last edited: