Jump to content


Why PowerPC is done


  • Please log in to reply
191 replies to this topic

#41 Dark_Archon

Dark_Archon

    Master Blaster

  • Members
  • PipPipPipPipPipPip
  • 1792 posts
  • Location:Rochester, NY

Posted 19 July 2007 - 08:44 PM

View PostFendersrule, on July 19th 2007, 03:14 AM, said:

I have a 20" iMac, and it' s a great machine. It eats my G5 in OSX "zippyness". But for games, my G5 has the clear upper hand with it's 6800 Ultra. The iMac is also a $1800 machine, which tends to lead me to believe that most of the mac-buyers will buy something cheaper like a mini or Macbook...which can't play games worth the asshole.

I was under the impression that Dual G5s were still faster than the mobile Core 2 Duos... I remember seeing that they were much faster than the mobile Core Duos(Yonah?).
Mac Pro 2.66 Ghz NVIDIA GeForce 8800 GT 7 GB RAM SONY DW-D150A SuperDrive

#42 Huntn

Huntn

    Verbal Windbag

  • Members
  • PipPipPipPipPipPipPipPipPip
  • 4074 posts
  • Pro Member:Yes

Posted 19 July 2007 - 09:30 PM

View PostSmoke_Tetsu, on July 19th 2007, 08:53 PM, said:

Point is people like me don't need to be told when to give up on certain systems. It's not like I'm sitting here waiting for some random stranger on the internet to tell me to upgrade. ;) :P

I think it becomes self evident when you hardware can no longer hack it. :)

#43 Quicksilver

Quicksilver

    Verbal Windbag

  • IMG Writers
  • 4227 posts
  • Location:Chicago Illinois
  • Pro Member:Yes

Posted 19 July 2007 - 10:01 PM

View PostDark_Archon, on July 19th 2007, 09:44 PM, said:

I was under the impression that Dual G5s were still faster than the mobile Core 2 Duos... I remember seeing that they were much faster than the mobile Core Duos(Yonah?).

Actually, they're not.  Core for core, a single Core Duo core is just about the same speed as a single G5 core.  Macworld did a test a while back where they disabled one of the cores in a 2 GHz CD iMac and ran it against a 2.1 GHz G5 iMac, and over a series of a half dozen tests, performance was pretty much the same (the G5 had a slight edge).
Former Senior Hardware Editor
InsideMacGames.com

#44 Smoke_Tetsu

Smoke_Tetsu

    Uberspewer

  • Members
  • PipPipPipPipPipPipPip
  • 3318 posts
  • Steam Name:Tetsu Jutsu
  • Steam ID:smoke_tetsu
  • Location:Cyberspace

Posted 19 July 2007 - 10:08 PM

Another thought is even if they are equal or the G5 is faster if the games don't have the same kind of optimizations on both they won't be the same speed on both.
--Tetsuo

Alex Delarg, A Clockwork Orange said:

It's funny how the colors of the real world only seem really real when you viddy them on the screen.

the Battle Cat said:

Slower and faster? I'm sorry to hear such good news?

Late 2012 27 inch iMac, Core i7 Quad 3.4GHz, 16GB RAM, Nvidia GeForce GTX 680MX 2GB, 3TB HDD - Mavericks

Late 2009 27 inch iMac, Core i5 2.6GHz, 12GB RAM, ATI Radeon 4850HD 512MB, 1TB HDD - Mavericks

Mac Mini, PowerPC G4 1.4Ghz, 1GB RAM, Radeon 9200 32MB, 256GB HDD - Leopard

Dell Inspiron 1200 Notebook: 1.2GHz Celeron, 1.2GB RAM, Intel GMA915, 75GB HDD - Ubuntu

Generic Black Tower PC, Dual Core 64-bit 2.4GHz, 4GB RAM, GeForce 9600 GT 512MB - Windows 7


#45 rbarris

rbarris

    Valve Software

  • Developer
  • PipPipPip
  • 219 posts
  • Location:Irvine CA

Posted 20 July 2007 - 12:54 AM

View PostQuicksilver, on July 20th 2007, 04:01 AM, said:

Actually, they're not.  Core for core, a single Core Duo core is just about the same speed as a single G5 core.  Macworld did a test a while back where they disabled one of the cores in a 2 GHz CD iMac and ran it against a 2.1 GHz G5 iMac, and over a series of a half dozen tests, performance was pretty much the same (the G5 had a slight edge).

I wouldn't mind seeing a link to those tests; I have a hunch they weren't games.  It would be difficult to factor out the difference between the X600 in the iMac PPC and the X1600 in the iMac Intel.

The CoreDuo has four times the L2 cache and half the memory latency of the G5 in the iMac.  I have a hard time seeing how anything not disk bound would put those two parts close together in rating, on a game-like workload.

Suppose you could do another version of that test where you only allowed the G5 to use half of its registers since it has so many compared to CoreDuo... you might skew the test in an interesting direction but it doesn't reflect the real end user experience where they will use all of the machine's resources if possible.

#46 Quicksilver

Quicksilver

    Verbal Windbag

  • IMG Writers
  • 4227 posts
  • Location:Chicago Illinois
  • Pro Member:Yes

Posted 20 July 2007 - 01:08 AM

Here's the article, dating back to early 2006.  They separated the single and dual processor charts, so you'll have to do a lot of scrolling back and forth.  In my opinion, this article is one of the few decent Macworld articles released in the past eight years--my dad had a subscription to Macworld from the first issue in 1984 until late 1999, when I (who was the only one in the family that had time to read it) told him that it was time to cancel the subscription because I felt like they were becoming decidedly content light and clearly biased toward Apple.  Turns out I was correct, if you read that article about Colin Crawford (recent CEO of Macworld).
Former Senior Hardware Editor
InsideMacGames.com

#47 Fendersrule

Fendersrule

    Fanatic

  • Members
  • PipPip
  • 130 posts
  • Location:Central Washington

Posted 20 July 2007 - 01:15 AM

View Postrbarris, on July 20th 2007, 01:54 AM, said:

I wouldn't mind seeing a link to those tests; I have a hunch they weren't games.  It would be difficult to factor out the difference between the X600 in the iMac PPC and the X1600 in the iMac Intel.

The CoreDuo has four times the L2 cache and half the memory latency of the G5 in the iMac.  I have a hard time seeing how anything not disk bound would put those two parts close together in rating, on a game-like workload.

Suppose you could do another version of that test where you only allowed the G5 to use half of its registers since it has so many compared to CoreDuo... you might skew the test in an interesting direction but it doesn't reflect the real end user experience where they will use all of the machine's resources if possible.

The Core2Duo has way more cache, and you are correct on memory latency....but have you ever seen the bus speeds on a G5?
2.2GHz MBP 7200 RPM
2.8GHz MP 1.5TB RAID & 8800 GT
1.5GHz G4 Mini
23" Cinema Display

#48 teflon

teflon

    Bastard of the Popeye Analogy

  • Members
  • PipPipPipPipPipPipPipPipPipPipPipPip
  • 9589 posts
  • Location:London, UK

Posted 20 July 2007 - 03:56 AM

also, the other advantage that the CD and C2D had over the G5 (despite being at a similar speed with a single core) was that it was less power hungry and hot, thus allowing its use in laptops. Add to that that theyre all (pretty much) dual core, and that means that you get more power for your buck.

especially in upcoming games.

there is no way that a G5 iMac will be able to play UE3 games, but any intel iMac with a GPU or MBP will be able to. Not only is it dual core, but its got a semi decent GPU in it too. Only a PM G5 (dual cores and above 2.0Ghz) would really be able to stand up to that, and even then only with an upgraded GPU (preferably a 6800 or X1900, depending on the logic board).
Polytetrafluoroethylene to my friends.

Macbook Pro - C2D 2.4Ghz / 4GB RAM / Samsung 830 256GB SSD / Geforce 8600M GT 256Mb / 15.4"
Cube - G4 1.7Ghz 7448 / 1.5GB RAM / Samsung Spinpoint 250GB / Geforce 6200 256Mb
Self-built PC - C2Q Q8300 2.5Ghz / 4GB RAM / Samsung 830 256GB SSD / Radeon 7850 OC 1GB / W7 x64
and a beautiful HP LP2475w 24" H-IPS monitor

#49 Fendersrule

Fendersrule

    Fanatic

  • Members
  • PipPip
  • 130 posts
  • Location:Central Washington

Posted 20 July 2007 - 04:50 AM

G5 2.0GHz, 5GB RAM, and a 6800 Ultra here.
2.2GHz MBP 7200 RPM
2.8GHz MP 1.5TB RAID & 8800 GT
1.5GHz G4 Mini
23" Cinema Display

#50 rbarris

rbarris

    Valve Software

  • Developer
  • PipPipPip
  • 219 posts
  • Location:Irvine CA

Posted 20 July 2007 - 12:47 PM

View PostFendersrule, on July 20th 2007, 07:15 AM, said:

The Core2Duo has way more cache, and you are correct on memory latency....but have you ever seen the bus speeds on a G5?

The bummer about the G5 bus wasn't the clock rate.  The towers ranged from 600-1250 MHz, but the iMacs never went over 600-700 IIRC.

The bummer is in the northbridge to memory, which was nowhere near as fast as Intel's, and the 32-bit-wide read path.  Compare with the Intel bus where you can read at 64b wide.

#51 teflon

teflon

    Bastard of the Popeye Analogy

  • Members
  • PipPipPipPipPipPipPipPipPipPipPipPip
  • 9589 posts
  • Location:London, UK

Posted 20 July 2007 - 02:40 PM

the PMG5 tower's clock speed was always exactly half that of the processor's clock speed. The iMac's was always 1/3rd of the CPU's speed. So, in theory (with fast enough RAM), the processors would have been completely unhindered by access time slow down, while the iMac's would have been slightly hindered.

but as you point out, that didnt quite work out properly because of the northbride.
Polytetrafluoroethylene to my friends.

Macbook Pro - C2D 2.4Ghz / 4GB RAM / Samsung 830 256GB SSD / Geforce 8600M GT 256Mb / 15.4"
Cube - G4 1.7Ghz 7448 / 1.5GB RAM / Samsung Spinpoint 250GB / Geforce 6200 256Mb
Self-built PC - C2Q Q8300 2.5Ghz / 4GB RAM / Samsung 830 256GB SSD / Radeon 7850 OC 1GB / W7 x64
and a beautiful HP LP2475w 24" H-IPS monitor

#52 XxtraLarGe

XxtraLarGe

    Master Blaster

  • Members
  • PipPipPipPipPipPip
  • 2324 posts
  • Location:Michigan

Posted 20 July 2007 - 03:06 PM

Say what you will about PPC, but Civ IV & WoW run fine on my PowerBook G4. Of course, that's because Aspyr & Blizzard both rock as far as Mac support goes.

#53 Eric5h5

Eric5h5

    Minion Tormentor

  • Members
  • PipPipPipPipPipPipPipPipPipPipPipPip
  • 7160 posts

Posted 20 July 2007 - 03:47 PM

View PostXxtraLarGe, on July 20th 2007, 05:06 PM, said:

Say what you will about PPC, but Civ IV & WoW run fine on my PowerBook G4. Of course, that's because Aspyr & Blizzard both rock as far as Mac support goes.

WoW is probably fine (haven't played it), but Civ IV is definitely not "fine" even on my 2.5GHz G5 + X800.  Not what I consider fine anyway.  All other games are, though (Quake IV runs great, etc.).

--Eric

#54 Quicksilver

Quicksilver

    Verbal Windbag

  • IMG Writers
  • 4227 posts
  • Location:Chicago Illinois
  • Pro Member:Yes

Posted 20 July 2007 - 03:59 PM

I'd say that WoW is "fair" on a PowerBook G4.  I've seen it running on quite a few PowerBook G4s at my school's ACM office, and they can't crank it up to full texture quality at native resolution without serious performance problems.
Former Senior Hardware Editor
InsideMacGames.com

#55 teflon

teflon

    Bastard of the Popeye Analogy

  • Members
  • PipPipPipPipPipPipPipPipPipPipPipPip
  • 9589 posts
  • Location:London, UK

Posted 20 July 2007 - 04:47 PM

but can you run it at native resolution with medium texture quality? after all, the PB G4 had 64mb of VRAM IIRC and thats the main factor in in the size of textures and various mappings (bump maps, specular maps etc. etc.) that you can run.
Polytetrafluoroethylene to my friends.

Macbook Pro - C2D 2.4Ghz / 4GB RAM / Samsung 830 256GB SSD / Geforce 8600M GT 256Mb / 15.4"
Cube - G4 1.7Ghz 7448 / 1.5GB RAM / Samsung Spinpoint 250GB / Geforce 6200 256Mb
Self-built PC - C2Q Q8300 2.5Ghz / 4GB RAM / Samsung 830 256GB SSD / Radeon 7850 OC 1GB / W7 x64
and a beautiful HP LP2475w 24" H-IPS monitor

#56 Rev-O

Rev-O

    BANNED

  • Members
  • PipPipPipPipPip
  • 1215 posts
  • Location:Colorado

Posted 20 July 2007 - 06:50 PM

Just from a playing experience I can say that Civ4 on my MBP (2.33 C2D 3gb ram 256mb X1600) feels way more snappy than Civ4 on my G5 PowerMac (dual 2Ghz, 3.5gb ram, 6800 ultra). The resolutions were different (17" MBP vs 30" cinema display), but the performance on the G5 made me stop playing Civ4 (much boggage about halfway through a huge map with a few civs) and Civ4 is a charm on the MBP. Same for Doom 3. Feels way snappier on the MBP.
In day to day computing my MBP feels peppier. Very unscientific, I know.

#57 Quicksilver

Quicksilver

    Verbal Windbag

  • IMG Writers
  • 4227 posts
  • Location:Chicago Illinois
  • Pro Member:Yes

Posted 20 July 2007 - 07:22 PM

View Postteflon, on July 20th 2007, 05:47 PM, said:

but can you run it at native resolution with medium texture quality? after all, the PB G4 had 64mb of VRAM IIRC and thats the main factor in in the size of textures and various mappings (bump maps, specular maps etc. etc.) that you can run.

Yeah, you can--but even the 128MB versions had trouble with WoW at medium-to-high quality.  You have to remember that the fastest PowerBook G4 (the 1.67 GHz model) only managed 28.4 fps in the (original) Unreal Tournament Antalus botmatch at 1024x768.  Those things were incredibly underpowered, imho.

View PostRev-O, on July 20th 2007, 07:50 PM, said:

Just from a playing experience I can say that Civ4 on my MBP (2.33 C2D 3gb ram 256mb X1600) feels way more snappy than Civ4 on my G5 PowerMac (dual 2Ghz, 3.5gb ram, 6800 ultra). The resolutions were different (17" MBP vs 30" cinema display), but the performance on the G5 made me stop playing Civ4 (much boggage about halfway through a huge map with a few civs) and Civ4 is a charm on the MBP. Same for Doom 3. Feels way snappier on the MBP.
In day to day computing my MBP feels peppier. Very unscientific, I know.

Well, it is faster.  I had a similar system to yours (Dual 2.0 GHz, 2 GB RAM, X800XT), and it had trouble with Call of Duty 2 on high settings (not "very high") at 1024x768.  This MacBook Pro has no problems running Half Life 2: Episode One at 1440x900, all settings maxed, with 4x FSAA and 16x anisotropic filtering enabled.  Of course, we're talking about a game on Mac OS X and a game on Windows XP, but I think that the fact that HL2:E1 is HDR-enabled, a Shader Model 3.0 game, and running on a laptop evens things out a bit.  ;)

Obviously, using Intel-standard stuff really helps us as gamers.  I don't really trust Apple when it comes to feature implementation--obviously, they're freakin' amazing at everything they do, especially hardware design.  It's just that they don't have an interest in providing support for things that they don't care about.  I seem to remember that the G4s (and perhaps the G5s) lacked basic AGP write combining--one of the easiest and well-known ways to speed up GPU performance.  Because of that (and the old 100-167 MHz main bus), you couldn't push Halo's timedemo scores past 30 fps on a G4 at 1024x768 regardless of video card, because the processor simply couldn't feed enough data to the GPU.
Former Senior Hardware Editor
InsideMacGames.com

#58 Thinine

Thinine

    Fanatic

  • Members
  • PipPip
  • 104 posts

Posted 20 July 2007 - 07:53 PM

View PostRev-O, on July 20th 2007, 07:50 PM, said:

Just from a playing experience I can say that Civ4 on my MBP (2.33 C2D 3gb ram 256mb X1600) feels way more snappy than Civ4 on my G5 PowerMac (dual 2Ghz, 3.5gb ram, 6800 ultra). The resolutions were different (17" MBP vs 30" cinema display), but the performance on the G5 made me stop playing Civ4 (much boggage about halfway through a huge map with a few civs) and Civ4 is a charm on the MBP. Same for Doom 3. Feels way snappier on the MBP.
In day to day computing my MBP feels peppier. Very unscientific, I know.
:blink:
Of course it's slower, you're playing at 4x the resolution. At the same resolution I'm pretty sure the G5 would be a bit faster.

#59 Huntn

Huntn

    Verbal Windbag

  • Members
  • PipPipPipPipPipPipPipPipPip
  • 4074 posts
  • Pro Member:Yes

Posted 20 July 2007 - 09:43 PM

View PostQuicksilver, on July 20th 2007, 04:59 PM, said:

I'd say that WoW is "fair" on a PowerBook G4.  I've seen it running on quite a few PowerBook G4s at my school's ACM office, and they can't crank it up to full texture quality at native resolution without serious performance problems.

Which powerbook? Did not see it in your sig. Wow runs good on mine. Speaking of signatures, how is Nexus, alot like X2 or better?

#60 Quicksilver

Quicksilver

    Verbal Windbag

  • IMG Writers
  • 4227 posts
  • Location:Chicago Illinois
  • Pro Member:Yes

Posted 20 July 2007 - 11:57 PM

I meant a 1-1.5 GHz PowerBook G4 with a 64 MB Radeon 9700 Mobility.

Nexus and X2 are completely different types of games.  X2: The Threat is sandbox space sim, played from the first-person perspective with a huge emphasis on commodity trading.  It's a good game if you like being a space banker. ;)  Really, though, the thing that drives me crazy about Egosoft's X-series is that it teases you with its potential.  Basically, they have a great engine, but very little focus and writing ability.  If they kept the X3 engine and spent a few million dollars trimming down the gameplay, writing a blockbuster script, and implementing it, it'd probably knock a the blocks off a lot of people.

Nexus: The Jupiter Incident (on sale for $10 on Steam right now--a mind-blowing steal!) is an action space sim/RPG with a heavy emphasis on capital ship-to-ship combat.  If you've played Homeworld, I'd have to say that Nexus' story is excellent, but it isn't as as epic and cool as Homeworld's story (which was amazing).  The combat, on the other hand, is a lot more fun since you're a lot more hands on.  I'm a little past 3/4 of the way through the game, and I'd rate the game 8.0 out of 10.0 based on what I've seen.  The three things that keep this game from approaching the ~9.0 range are a mildly confusing ship configuration screen (there's a bewildering array of engines, missiles, lasers, cannons, sensors, and other cool gadgets to install in your ship), slightly outdated (circa 2004) graphics, and an annoying figure that lasts for 3-4 levels.  Fortunately, he gets his ass handed to him in a satisfying way, so it's actually a part that I might even look forward to when I play the game again.  Before I forget, Nexus' audio is excellent.

Anyway, if you know the way I review games and hardware, 8.0 is an excellent score (I really use the full 0-10 range).  If you want more information on the game, I'd be happy to go into more detail.  Regardless, if you loved Homeworld, this is a must have game for you--it'll be the best purchase you've made in a while.
Former Senior Hardware Editor
InsideMacGames.com