Gaming Resolution Scaling
Taking a further look at gaming performance at other resolutions, here are the same games from the previous page with additional results at 1440x900 and 1680x1050 for the G50V and dv5t - we didn't feel there was any need to attempt to run U6V at these higher resolutions. We used an external LCD for the dv5t, but note that you can purchase the laptop with a WSXGA+ display so the results there are actually meaningful. We included the Gateway M-152XL and P-7811 as points of reference; referring back to our midrange trio of laptops, the M-152XL was slightly faster with its HD 2600 than the 8600M GT/9500M GS, and of course the P-7811 with its 9800M GTS is easily the fastest sub-$1500 gaming notebook on the market.
The 60% average performance lead of the P-7811 over the G50V jumps up to a 77% lead at 1680x1050, and quite a few titles will require lower detail settings to get acceptable performance. The dv5t with 9600M GT is significantly slower than the G50V at higher resolutions; the P-7811 is 2.6 times as fast as the dv5t at 1680x1050 and the G50V is about 50% faster.
Facebook
Twitter
RSS
27 Comments
View All Comments
JarredWalton - Friday, October 24, 2008 - link
I chatted a bunch with ASUS on this; there was some confusion so I may have ended up with the wrong conclusion. (Yeah, marketing wasn't positive on the specs, and engineering didn't ever pass on the exact details.) I actually had a paragraph detailing the differences between the 9800M GTS and this supposed 9800M GS. Since I don't have one in my hands, I can't say one way or the other with certainty.The worst case would appear to be clock speeds equal to that of the 8800M GTS (500 core instead of 600 core on the 9800M GTS), which is still going to be a lot faster than these other notebooks. Since it's also limited to 1366x768, gaming performance should be no problem at native res... but there's a lot of headroom left untapped. Certainly, gaming performance won't be lower than the G50V tested here, unless the game happens to be CPU limited.
Enrox - Friday, October 24, 2008 - link
Take a look at the Gateway P-7811 battery's life: it's about 150 minutes regarless the task (DVD playback, web surfing, H.264 playback).That to me says only one thing: no power management in place.
Is that a Vista issue or a BIOS issue?
JarredWalton - Friday, October 24, 2008 - link
Oh, the P-7811 is definitely doing *something* - though idle battery life is lower than I'd expect relative to the others. Actually, I think it's more that the P-7811 is doing quite well in other tasks. Remember: 17" 1920x1200 LCD, 7200 RPM HDD, and a 9800M GTS put it at a much higher power envelope than most of the other laptops. Relative to the P-6831 and m15x, the results seem to be right where you'd expect. If only Gateway had implemented Hybrid Power....jonmcc33 - Friday, October 24, 2008 - link
Did you verify that with the Power Saver setting that EIST was working properly? Use CPU-Z or similar to see if the clock speed of the FSB and CPU does change as it should. Check the BIOS settings as well.CU - Friday, October 24, 2008 - link
Yes it would be interesting to know what the cpu, gpu, fsb, and ram clocks are at when in power saving mode for Vista and OSX.JarredWalton - Friday, October 24, 2008 - link
CPU speed drops to a 6X multiplier, so at least that aspect is working. Looking at the voltages (according to CPU-Z), they're all at 1.083V except for the G50V, which runs at 1.338V most of the time. (I'm still trying to figure out what's going on there and will update when I know more.) I'm not as concerned with G50V battery life, though, since it's in a different class of performance and size; it's the U6V and similar notebooks that need to do a lot better.Regarding RAM, GPU, and FSB, the FSB stays locked at the base speed - 1066 MHz on the Centrino 2 notebooks. RAM likewise stays at a set speed, in this case 800 MHz. 2D GPU clocks (according to GPU-Z) are 169 MHz core, 200 MHz (100 base) VRAM on all three of these notebooks. GPU-Z also reports a memory clock of 800 MHz (400 base) for the HP dv5t, which seems wrong - I though the 9600M GT was supposed to be much faster RAM, but apparently not.
JarredWalton - Friday, October 24, 2008 - link
Update: The G50V was back on "High Performance" mode after rebooting (an issue with some of the ASUS software). Setting it back to "Balanced" or "Power Saver" dropped the CPU voltage to the expected voltage - though still slightly higher than the others at 1.063V.fabarati - Friday, October 24, 2008 - link
Asus is known for their crappy batterylife in the latest generation. When compared to equal or even better specced laptops, they fall flat on the ground. It's probably because of bad ACPI coding. My F8Sa has worse battery life than my old A8Js, despite having less powerhungry parts. And the A8Js had mediocre batterylife (I reached about 3½ hours, with hardware disabled). I can barely break 2 hours, and that's when I disable hardware.The HP DV5 seems to suffer from the same issue, at least that's the conclusion we came to when it was tested by NBR.
JarredWalton - Friday, October 24, 2008 - link
I haven't tested a comprehensive selection of laptops by any means, but if you look at the specs for the various laptops and the resulting Minutes/Whr chart you can see that if this is bad ACPI coding the practice extends far beyond just ASUS and HP. If the MacBook Pro was around 3 or 4 Min/Whr, I'd think maybe it was just some fine tuning that was missing, but it's still literally double what the closest tested Vista laptop managed.The best result I've personally seen on Vista to date is the http://www.anandtech.com/mobile/showdoc.aspx?i=328...">ASUS U2E, which manages 3.72 Min/Whr with the 86.5 Whr battery. That's a lot closer than the other laptops, but keep in mind that has a U7500 CPU (10W max TDP), X3100 IGP, and an SSD, plus an 11.1" LED LCD.
Another 15.4" laptop I'm currently testing with T7250 and X4500 graphics (plus 4GB RAM, 250GB 5400RPM HDD) manages 4.18 Min/Whr, which is closer to Apple. Still, that's a 50% advantage for the MacBook, so it's not really *that* close. (It gets 204 minutes of battery life in our web surfing test.)
nizanh - Friday, October 24, 2008 - link
Can't you just install Vista on one of the MacBooks?Sounds to me like the best testing methodology.