The Compal IFL90 (PowerPro J 10:15) is now available for PRE-ORDER with SHIPPING BEGINNING MID-JUNE

The IFL90 fully supports Windows XP 32/64-Bit.

At this time we are under the assumption one of the cards will be DDR3. We hope to have confirmed information in the coming days (including the clocking)

 

What does this mean?

 

GDDR3 is a different kind of RAM module that they put on graphics cards. It is closely related to DDR2 for system ram. It uses less power and runs a bit cooler which means it can be easily overclocked.

There is also GDDR2 which is a little bit cheaper to produce, so you can still find some cards that use this dated technology. However, more than likely, you will not notice a performance change if you're using GDDR2 or GDDR3, real world performance I mean (and even with OCing it, I doubt you'd see that much of a difference).

 

deserteagle - does 3DMark06 constitute real world performance (because initial tests show a 20-25% performance difference in that benchmark)?

Petrov.

 

Petrov, 3DMark06 is the antithesis to real world performance.

Synthetic benchmarks can be frustrating exactly because they are NOT real world performance, and yet they can throw out measurements such as that which you have quoted which for real world performance are virtually meaningless.

 

Thanks paladin - would you expect the fps difference between GDDR2 and GDDR3 to be substantially less than 20-25% on most graphically intensive current games then (at, say 1680x1050)?

Petrov.

 

Yes, the difference would be substantially less than the 20-25% measured by 3DMark06.

For reassurance please see this post...and click on the link in it:
http://forum.notebookreview.com/showpost.php?p=2014696&postcount=108

 

Yup, read this - thanks again, appreciate the quick response as always. I look forward to seeing some more real world benchmarks come out!

Note, the Zepto IFL90 has GDDR2: http://www.zeptoforum.com/viewtopic.php?p=4164#4164

Petrov.

 

Hey Donald, you should start some kind of sticky to campaign the over-use of synthetic benchmarks since they are misleading to the unknowing readers and often have little bearing on the user experience.

 

I don't control the Sticky's, but I agree that someone should start a thread explaining the un-value of synthetic benchmarks.

Better for a power user to do it than a dealer though (hint..hint )

 

I'll look into it after work, nice subtlety

 

Thanks pyro...I'm sure we will all appreciate your effort, particularly knowing your expertise.

 

I'm probably going to try and check around before I start writing something up... I'm betting that something like that is more generic and wouldn't belong in just the Compal forum (I know the poor sager/clevo guys need it ) but I honestly don't see many people reading it if it's not local to "the name brand" they own.

 

if you could write some, I could add it in the compal faq.. although it probably would best be served in the video card section where everyone seems to sweat day and night for 3dmark numbers.

 

What would you guys recommend to run for an objective comparison of real-world graphical performance of the new laptops (IFL90, Zepto's, G1S, etc)?

Petrov.

 

Comparing the same game, on the same map/zone/level, with the same options , internet connection (if applicable) and driver versions is about as good as you can get if you really want to compare..

the only problem is that the people who sweat about justifying their grand purchase don't like to hear that their system only get's a few FPS more then the guy who paid $$$ less in the real world.. What they want to see is the 1.5K more 3Dmarks..

The reason I say to use a real game, is because different games use different game engines, and have cleaner/dirtier code. Some are also better suited for ATI, while some are more suited for Nvidia. Using a real game you intend to play is the best measurement because it doesn't like using a prime engine like in 3Dmark.It also lets you see how it will play at the actual settings you want to use.. Most games honestly look fine at < Native Resolution with the right settings, and they will always have higher FPS at lower resolution.

This isn't to say synthetic test's are meaningless.. The problem is that there isn't much environment controls for the testing, and the number generated isn't a reflection of real world, it's a reflection of potential, so it's only really good for comparing to other cards that were also tested on the same software. This leads to the lack of controlled hardware though, so the numbers are always skewed.

As an example, I've seen a driver version change make hundreds of points different 3D mark scores, yet in real life the difference is only 1-3fps..

 

Real games:

Doom 3 and Fear are pretty good games for benchmarking. Ofcourse this is only directx9. Supreme Commander also is a good game because it puts high demand on the non GPU parts of the laptop as well.

Company of Heroes with the directx10 patch might be ok for a directx10 test for now.

 

Fear and Supreme Commander are excellent, Unreal Games also have great engines, Anything from ID software is reliable... I'd say the torque engine, but that's a bit outdated.. (Tribes games)

 

Thanks for the support Goren
Maybe sticky on the Video Card section, and add reference in the FAQ section for the different brands?

 

Isn't the problem that GPU drivers can be written to exploit certain scenarios in synthetic benchmarking programs? Could a more generic benchmarking program be written using say 3 different gaming engines and averaging the results, or would that still not be real world enough?

It would also help if the reporting included GPU and driver information so you can make an apples to apples comparison.

 

You would also need to report HDD size and speed (and manufacturer should the platter's be different sizes), CPU Type/speed, Ram used..

 

Sorry for this little tangent, but on the subject of HDD's, I'm looking for a solid opinion on what HDD to get with an IFL-90, I know the 7200rpm Seagate will be faster but louder and hotter (I'm guessing) so since I don't do a whole lot of gaming, would I be better off with a 5400rpm HDD or the 7200rpm?

Thanks!

 

The 160GB 5400RPM disk is pretty fast. The hard disk wont effect gaming much anyway, aslong as you got 2GB or ram or more. Mostly loading times and start up only.

 

Oh where are the hybrid 200GB 7200RPM drives?

 

Google the Seagate 7200.2 drives and read the reviews. There are also some good treads on them if you search this site. It doesn't sound like these drives are much hotter or louder than 5400rpm drives. They do consume alittle more power, however.

After reading about them I decided on the 120Gb 7200.2 drive.

 

I think my point is that it should be possible to design a level benchmarking program that more closely matches real world performance using gaming engines and maybe some pieces from OpenOffice to build a business oriented benchmark. It should report all the information you mention to make sure you're getting a good comparison between machines.

 

I get what you mean, In theory it sounds good, but I'm sure if it were that easy, it would already be done

 

If you don't have a real need (which this makes me think you dont) I'd save the money, or upgrade a larger drive. Shave yourself a few minutes of battery life too maybe.

The real speed comes down to the platter size ratio to the RPM, because a larger platter on a faster drive, may actually be slower then a smaller platter on a slower drive.

 

On the IF90.
Is the 8600GT 512MB RAM GDDR2 or GDDR3.

 

sounds good. I'll let Chaz and the other gpu section mods know about it once you make it, then I'll link it in the compal faq.. maybe the Asus one too if Cori lets me

 

We are still waiting on a full confirmation of that as well as the GS or GT version of the card, but the Zepto version of the IFL90 in Germany is GDDR2.

 

Are we sure the Zepto is based on the Compal? The old Zepto wasn't based on the HEL-80, was it?

 

The 3415 model is based off the IFL90, based on this thread

I don't know much beyond what was discussed (literally)

Edit- The link to the German forum posted at the end by Petrov is where the confirmation on the GDDR2 came from.

 

They've released a new model just a few days ago that is built off of the IFL90.
As far as the Znote 6625W goes, I'm not sure if that is or not, either way though it has similar specs.

Edit: Dang it sco-fri beat me to it.

 

I'm thinking the guys over in the sager clevo room need, all those D900C and 570u/ru people are starting to twitch

 

To be fair, if I could get a D900c, I'd twitch too

 

6625W is Inventec ODM. =)

 

No, no, no! A larger plater is faster than a smaller one of the same speed, since the heads have to travel less distance to reach their target. It comes down to memory density. If the platters are the same, but there are more of them, there will not be an increase in speed.

 

Anyone have any estimates on how long it'd take to get the laptop if i order after the final word on the gfx card comes out? As in I'd be after all the preorders

 

Please define larger:

Larger = physical size,

OR,

Larger = more capacity

 

There are already several hundred pre-orders in, so at this point it would be quite difficult to estimate how long it would take to get the laptop if you wait to order it. How soon do you need it?

 

What you and I said are the same thing..

120GB 7200RPM drive using 60GB per side platter will be potentially slower then an 80GB 5400RPM drive using a single sided platter since the platter on the larger/faster drive has a lower density. (this is what you quoted me for being wrong about)

Not only is the density a concern, but is is the number of r/w heads, how many actuators are used to hold the heads (single fixed etc). The density also adds to the record index size required, so a high dense platter will actually start to slow down because the index size takes to long to dig through. If you don't have enough read heads to compensate the index size, or your drive design shares heads across r/w functions, you can slow down the performance. This complication is why notebook drives have taken so long to catch up to their desktop counterparts.

I worked at Seagate for a bit, I know what I'm talking about here ..

 

uhmmm hopefully by mid-July haha. I'm going to college this fall and I'm just starting to look at laptops. I'll probably be ordering whatever I get in the next 2-3 weeks

 

So how many platter does a 100GB 7200rpm Seagate hard drive have?
Tim

 

Wow! Several hundred? Do you expect the supply to be adequate or will we have to wait a while?

 

ditto. that's what i was basically wondering haha

 

looks like it will be a long wait... =/

 

I hope not. I have a place in line someplace - where I don't know. A few hundered from one dealer out of all of them world wide seems like a lot though.

 

I don't know specific models, it's been years since I've been at Seagate, but something makes me think it would only have one platter with 100GB per side, or 50 per side. I was in the desktop parts area, but the technology is the same.

If you gave me a more specific name I could find out.

 

I am not saying that PowerNotebooks.com has several hundred pre-orders, that is for the US market.

The initial import order is for over 3,000 but they don't always all arrive in the first shipment. Therefore there is no way to predict exactly how many will arrive in the first shipment, but once mass production starts the shipments will be coming quite regularly. However trying to predict when someone will get theirs when they haven't even placed an order is quite impossible. Once the order is placed we can give a much better estimate since we will know at that time what the back order situation is, and how many come in during the first couple of weeks of mass production.

The best thing to do is not to be so anal about GT or GS (although we are pretty sure at this point that the WSXGA+ will be GT) and just decide what you want to do and go for it so you can secure your place in line.