Jump to content
Sign in to follow this  
asaggynoodle

*Almost* Fastest PowerBook ever, and more to go.

Recommended Posts

Close, so close, but it's not the fastest PowerBook ever, not yet (at least not from a benchmark standpoint).

 

https://68kmla.org/forums/index.php?/topic/21144-haplains-never-ending-quest/?p=263061

 

Sorry it's been a while since the last update, I screwed up the last motherboard and had to wait for a new one from Australia.. (Go figure).

I started screwing with the FSB, but it sucks because OS X doesn't read the current MaxBus frequency from the register. Only the stock one seems to get reported. 

 

TADA!

http://browser.primatelabs.com/geekbench2/2607788

 

There is still more to go since I figured out the Voltage Nets, I can go all the way up to 1.6v without doing anything exotic. 

This was achieved at a 172Mhz MaxBus with a 12x Multiplier.

The new motherboard I have wasn't stable at 2Ghz on stock voltage like the last one. I'm currently only at 1.4v, I already tested 1.6v out and it runs fine under load with my new NT-H1 thermal paste.

 

The next step is full blown 194Mhz MaxBus at a 12x multiplier. So 2328Mhz. It should be fine, since 1.6v is a good bump of vCore. I'll be doing this sometime soon. 

I could still play with the multiplier, but I got a little scared off when I took the board out and it didn't work at 12.5x. Turns out the DFS setting on the Motherboard doesn't support half steps... *Sigh*

Anything above 12x requires atleast 3, 0402 resistors to be moved on the bottom of the board. So no thank you.

 

I found it interesting even though the RAM and CPU are both Overclocked now, the Memory performance isn't improved.

 

So officially the fastest Powerbook Ever? Please?

Edited by asaggynoodle

Share this post


Link to post
Share on other sites

So officially the fastest Powerbook Ever? Please?

 

Nah, we're going to keep throwing into your face the fact that Apple *did* make a prototype dual G4 which hypothetically might have been just the teeniest bit faster forever. ;)

 

(I also know someone who *swears* to have once actually had their fingers on a G5 Powerbook prototype, but I've been taking that one with a "pictures or it didn't happen" grain of salt since the day I heard it. Not that I believe that would necessarily outrun your overclocks even if it was real, given the G5's spotty performance reputation.)

Share this post


Link to post
Share on other sites

I don't see why Apple wouldn't have made a mockup or a G5 laptop. IBM made a Thinkpad with a full blown desktop P4 and managed to cool it (G40/G41 series).

I've come to the conclusion apple intentionally crippled the PowerPC mobile options there at the end. You have to figure the Intel/Apple negotiations had to have been at least a year or two. That would have given plenty of time for Apple to come up with a road map to incorporate x86 into their lineup. By late 2005 there was no competitive reason not to have either a Dual G4, or some kind of "G5" solution. *Puts Tinfoil hat on* What I'm trying to say, is that I PERSONALLY believe that Apple had intentionally handicapped the mobile options so that the performance comparisons between the last G4 PowerBook, and the first MacBook with x86 looked much larger than they had actually been. I mean, from a marketing standpoint that is probably the best thing you could have ever done. That would boost sales as everyone wants "40% FASTER!" performance from the year before.

 

I might be completely wrong here, but I don't understand how you can have a nearly 3Ghz Quadcore PowerPC chip in your desktops that was performance competitive with any of the Intel Offerings. Then your highest performance Notebook is still running with a Single core chip at half the clock speed. That's just too big of a spread to make me believe they had been trying their hardest to compete.

 

On the Overclocking side of things, I don't think I'm going to be getting much higher. I increased the MaxBus to the next step which is only a 5Mhz bump (12x) so 60Mhz Clock Speed. That was NOT stable. I started getting Geekbench scores in the range of about ~600. Clearly something was not working right. Changed out all of the RAM, though it might be something there. Same problem persisted. So I bumped up the voltage to 1.60v and gave it a swiz. I couldn't even get it to boot up fully. No crash, no nothing, just would shut off. I noticed there was literally no heat, not even the fans came on before the shutdown. So I'd come to the conclusion the VRM's couldn't put out that much power when the Overshoot happened on the CPU. Went back, changed it down to 1.5v, and got it to boot to desktop, but it still was not stable from the power delivery point of view. I could go about 30-45 seconds and the machine would reproduce what had happened at 1.60v. Sometimes It would crash at the Apple logo during boot up, so same problem as before.

 

Looks like I've hit the limit on this board, nowhere near what was doable on the old one though. Seems to be very much luck of the draw in terms of the untapped potential on these chips.

Share this post


Link to post
Share on other sites

I don't see why Apple wouldn't have made a mockup or a G5 laptop. IBM made a Thinkpad with a full blown desktop P4 and managed to cool it (G40/G41 series).

 

The reason I'm a bit skeptical about the particular eyewitness account is they claimed to have had a G5 prototype that was a completely buttoned-up machine that looked "basically like a regular powerbook but thicker", not an obviously prototype thing with acrylic case parts, etc. Not to say it's impossible that they got as far as making fully-skinned machines, I'm sure Apple has the resources at their disposal to mill out some one-off aluminum enclosures, but... I dunno. The eyewitness in question is mostly reliable but has been known to fluff the truth now and again.

 

And, yeah, there were actually quite a few "laptops" that had full-blown P4s in them. For instance, Alienware made some gaming laptops that did have batteries but could run on them for about 30 minutes, tops. A guy at my company once pushed through an order for a few of them because he wanted laptops that could take 4GB of RAM when most systems topped out at two for a mobile VMware lab, and they were incredibly ridiculous things. I can understand why Apple wouldn't have felt comfortable slapping their name on something that ridiculous given the "stylish products" focus of the company. I'm sure there would have been Mac-using Pros willing to buy a brick if the performance would have justified it, but it would have been a *very* niche product.

Share this post


Link to post
Share on other sites

I would think Apple quit putting R&D into the G4 laptops because of the switch to x86 coming down the road. And if companies heard rumors of Apples switch to x86 they probably didn't bother with speed upgrades for the G4 and G5 chips. The dual core x86 chips would have killed any speed increase they could get for the G4 laptops anyway.

Share this post


Link to post
Share on other sites

I've come to the conclusion apple intentionally crippled the PowerPC mobile options there at the end. You have to figure the Intel/Apple negotiations had to have been at least a year or two. That would have given plenty of time for Apple to come up with a road map to incorporate x86 into their lineup. By late 2005 there was no competitive reason not to have either a Dual G4, or some kind of "G5" solution. *Puts Tinfoil hat on* What I'm trying to say, is that I PERSONALLY believe that Apple had intentionally handicapped the mobile options so that the performance comparisons between the last G4 PowerBook, and the first MacBook with x86 looked much larger than they had actually been. I mean, from a marketing standpoint that is probably the best thing you could have ever done. That would boost sales as everyone wants "40% FASTER!" performance from the year before.

 

I dunno. Intel's "Pentium M" chip was a real game-changer and Apple seems to have been caught really off-guard by it. The first Pentium M machines (which also included USB 2) came out about 6 months before the USB 2.0-equipped Aluminum Powerbooks, and I remember being shocked at just how well they performed compared to the 867mhz Titanium I had at the time. If you look up Geekbench results you can see just how badly the Pentium M *crushed* the G4 on a *clock-for-clock* basis. A January 2003 vintage Dell D600 with a 1.3ghz Banias Pentium M manages a score of 879; that's better than a 1.5ghz Powerbook a year and a half newer, and it manages it with a comparable power draw. While Apple certainly *could* have compensated by slapping in another G4 and going dual the resulting system would suffer accordingly in terms of weight, heat, battery life, etc.

 

Apple really was pretty boned here. The G5 didn't become remotely suitable for use in a laptop until the 970FX came out in 2004; to hit the sort of heat/battery profile Apple would have wanted it would have been limited at running in the very low 1ghz range at a time when Pentium M's were pushing up to the 2ghz mark, and PWRficient's Hail Mary attempt at an alternative also came a day late and a dollar short. (It's possibly worth noting that the Amiga X1000, which uses the one model of PWRficient CPU to actually see anything approaching widespread use and wasn't available until after the Intel transition was a done deal, actually seems to be slower than most of Apple's G5 desktops.) So... yeah, not sure I buy the conspiracy theory. I'm sure it would have been *possible* in some theoretical universe for Apple to have turned the tide by investing more in CPU development (IBM basically developed the G5 as a favor to them, only to have Apple repeatedly bite the hand trying to feed it), but Apple's not really a computer company, they're a consumer product company. Intel simply had more suitable off-the-shelf parts for what they wanted.

Edited by Gorgonops

Share this post


Link to post
Share on other sites

I wonder if down the road Apple will invest some of that extra cash into designing and building a custom desktop/laptop chip (or just buy AMD with change from the couch cushions). Cell Phone processors are getting pretty decent these days.

Share this post


Link to post
Share on other sites

Apple certainly does make more-than-decent ARM products, but I don't see them making a non-Intel "Mac" any time soon. I mean, let's be honest here, Apple hasn't exactly been investing a lot of effort in their computer line lately, and any ARM-based "Macintosh" would require both significant engineering effort and would pose issues with user education and expectations. (After all, Microsoft ended up confusing a lot of customers with the limitations of Windows RT vs. regular Windows regarding legacy software, and while Apple has a more positive track record with transparently handling the need for multi-platform code there is the issue here that unlike previous CPU migration efforts the ARM wouldn't be powerful enough to provide acceptable emulation for existing Intel binaries.) Not to say that Apple would never introduce some sort of "ARMbook", but if it happens it's far more likely to be an outgrowth of the IOS product family than a "Mac", per se.. Arguably it's already happened with the iPad Pro.

 

Really, when you consider the moribund state of the Mac Pro and how stale their computer line in general has become one has to wonder how long they're going to keep bothering.

Share this post


Link to post
Share on other sites

I think you're totally on to something here. I for sure thought the new MacBook was going to be ARM based. Looks like Intel managed to pull the Core-M out of their hats just in time. 

 

All Apple really needs (And everyone else) is a Co-Processor that is solely in charge of Static (Dynamic) Recompilation for x86-ARMHF at the OS level. That way they could bridge the gap between MacOS and iOS with no overhead. There are companies like those who made Eltechs Exagears, that can do native x86 run times on ARM at like ~96% of native performance. That more than likely wont ever happen, as that is a lot of effort. But it would be game changing, something apple needs (a game changer). Although similarly, Android/ChromeOS are basically becoming one. So it goes to show that with enough effort you CAN break down these barriers (Not the same, but still a big deal). At the end of the day, if ARM keeps going the way they are, it may be a matter of years before Apple does even more with ARM besides phones/tablets. If anything, looking at the iPad Pro starts to show the need for x86 level professional software, as the Hardware is already more than there. People (Real professionals) want to do real work on these platforms. 

Share this post


Link to post
Share on other sites

Core M's problem? Intel's tray price on it is $281 a pop. I'm sure Apple gets a better price, but I doubt it's *that* much better, and Apple doesn't have to make a profit on their A-series chips, which are estimated to cost $20-$40 a pop to manufacture. But I think any ARM move is five years away, at least. 

Share this post


Link to post
Share on other sites

I think when you have the buying power of Apple, you get some pretty deep discounts. You have to figure at least a 2.5x markup from BOM/Product Placement/R&D costs. I'd be willing to bet the whole Mainboard with all of the components probably costs around ~150-200$ per unit. The thing with ARM is that you have to pay the upfront cost of the IP Licensing which for a company like apple is probably around ~$10-$15 Million, then around 1-2% (of chip cost) in royalties per chip made. Talk about a Cash Cow. Although, when you have the volume to push product like Apple, I'm more than sure you can get them to sweeten the pot.

Share this post


Link to post
Share on other sites

Well, the thing is, that $281 is the tray price, not any kind of retail price. That's their price when you're buying them by the thousands. And Intel likes their margins just like Apple does. 

 

Even if they sell it to Apple for $120 a chip, that's still a good deal that can be saved. 

Share this post


Link to post
Share on other sites

I don't see why Apple wouldn't have made a mockup or a G5 laptop. IBM made a Thinkpad with a full blown desktop P4 and managed to cool it (G40/G41 series).

 

I used to have a Clevo/Jetbook (made by foxconn I think) laptop back in the day that had a desktop P4 2.0 in it. That thing would put off some serious heat. I think even some of the early Silver/Gray dells had desktop P4 chips. But they didnt last long. Market, or longevity. Motherboards failed alot. 

Edited by techknight

Share this post


Link to post
Share on other sites

Hell, even the laptop P4s were terrible in the thermal department. The P4m laptop I used to own got hot enough that when the power supply went nova and actually melted the plastic tip on the plug in the process, I didn't actually notice for about ten minutes.

Share this post


Link to post
Share on other sites
On 6/19/2016 at 11:21 PM, bunnspecial said:

I have very good reason to believe that the original poster does not in fact have a 7448 processor in his Powerbook as he claims.

 

Please see my full write-up on the subject at this link, and I certainly invite the OP to refute any statements which I've made.

 

http://forums.macrumors.com/threads/2ghz-powerbook.1977301/page-2#post-23036065

I find it rather odd that even all this time later, this has been so glossed over. OP clearly faked using a 7448. Reaching 2GHz on a 7447B isn't that hard, as proven by this iMac G4. Why @asaggynoodle felt like faking he had a 7448 is beyond me. Maybe to screw with the owner of EveryMac, as they posted his fake nonsense verbatim? Anyways, @dosdude1 has real 7448s, so we can finally see some non-faked 7448 action not too far from now ;) 

Edited by Daniël Oosterhuis

Share this post


Link to post
Share on other sites

CSB:

I worked at Freescale qualifying the 7448.  (And the 7447, 7450/55, 7410)

Apple kept Freescale on the hook for the 7448 right up to the bitter end, stringing FS along having them believe Apple was going to use the 7448.

 

At one point I was corresponding via email with an Apple engineer trying to get a firmware update for the P58 mobo's so I could run the Apple test suite and OSX with the 7448.

My email got forwarded to about three other people within Apple because they had already pulled the resources from the team that maintained their PPC test suite.

One of the Apple recipients replied with something like "Why do we need this? We're not using A8 (Apollo 8 ) and hit reply all so I was copied too.

 

My phone rang literally about 30 seconds after I received the email. It was the Apple engineer I had been corresponding with hastening to explain that I should ignore this particular reply and they were working to get me the firmware update I asked for.

I did get the OF update about three days later and about two weeks later FS got the "official" word that Apple would not be using the 7448. 

(Even though despite Apple's efforts it was apparent to FS long before that Apple wasn't going to use the A8. )

 

Anyway, I have some photos of the P58 (G4 MDD) test rigs I used at work. I should dig those out. They're hardly proprietary now.

/csb

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×