It matters, because games, especially as they get larger and more complex in the coming years (which is what 7 is designed for, rather than yesterday's software), will be dealing with larger data sets (memory management and file access), many more threads, to take advantage of multiple cpus (processor scheduling), etc.
A slow hard disk or fast hard disk will still see a performance improvement when used in a more intelligent manner. If you want to upgrade hardware, upgrade hardware. An OS is not a hardware replacement, and vice versa.
The idea of "levels" is retarded anyway. Fast disk i/o is needed for continuously streaming game worlds, which is where things are going.
Fair enough. The AMOUNT missing depends on hardware, granted. However, the fact that any ram is unusable at all, is due to the OS being 32 bit non PAE mode, which is a software issue. This was my point, i think we're referring to 2 different things...
Re:Does it matter, its all DirectX
on
Gaming On Windows 7
·
· Score: 4, Informative
DX is only part of the platform. DX doesn't cover stuff like file access, memory management, processor scheduling, etc...
It is noticeably faster than Vista. Its SMP support and responsiveness (as opposed to throughput, which I have not measured or compared) vs even XP is markedly improved.
If you have a single core box with less than 2gb, XP is probably as fast or faster.
If you have multiple cores, plentty of RAM (its CHEAP now, so if not why not), 7 will be quicker. Especially if you have a half recent 3d card, in which case much of the GUI is offloaded to it and its video memory...
Moving forward, the benefits of 7 over XP (or vista) will only become more apparent.
Same here. Well, with a small caveat. A couple of games that worked on vista do not work on 7 RC. namely (from my collection) : Vampire The Masquerade: Bloodlines, and Neverwinter Nights 2.
NWN2 barfs with a DirectX error (some missing string), No idea why VTMB dies (black screen, can only kill with ctrl+alt+del). Apparently the NWN2 problem is fixed in a build more recent than 7100, not sure about vampire.
But on the whole, if it works on Vista (and really, the jump from 98 to XP broke HEAPS more games than the jump from XP to Vista does), it will most likely work just fine on 7.
If you have need for more than 4GB RAM (not 3.5GB; that's the per-process virtual memory limit)
XP 32 bit can not use a full 4gb. It can only "see" 3.foo GB of ram, due to having only 32 bit address space, but needing to reserve some of that for purposes other than RAM. Eg, memory mapped IO.
Install 4GB ram into 3 different 32 bit XP boxes (hardware wise) and you'll likely have 3 different amounts of RAM available to the OS.
This is why we need to eliminate the oil, and build the engines using ceramics or teflon, or some other sort of friction-reduced/frictionless material.
Meanwhile, people have shit they want to get done (eg, streaming pirate hi-def video over their home network) and 802.11draft-N works just fine. Today. And has done for at least a year or two. I'm all for standards, but when it takes 3+ years for a standard to get a rubber stamp, when it works just fine already; something is wrong. From reading TFA, one of the hold ups mentioned was "acceptable co-existence language" mentioned in the draft. Sounds to me like the standards committee is broken.
*shrug* in the 5 years it's taken for the ratifying authority(ies) to pull their finger out, you could have purchased and de-commissioned draft-N gear.
As far as draft-N goes, my apple airport talks to a couple of Dell draft-N notebooks, a Thinkpad, my mac mini, and some china-spec draft-N card in my main home PC.
And in the meantime i've been able to actually stream high-def-ish content over 802.11n...
Back in the day, when I was about 15 and starting out trying to write demos (as in demoscene - we managed to replicate a few effects we'd seen back in the 90s from other big groups - after going "Wtf, how does that work?"), a friend and I started coding together.
The actual CODING is a one man job, sure; but when you're trying to solve a problem, having 2 minds work on the ALGORITHM gets you better results, quicker IMHO.
Sure, you'll possibly end up with the same algorithm/code eventually anyway, but having a second set of eyes and brain analyzing the process you're going to use picks up potential problems earlier. While you're thinking you're clever with some part of the problem, the other person is finding the glaringly obvious bug that you haven't seen (yet).
Old... but standard, available everywhere, and Good Enough (tm) for small scale, small team projects (some would argue, quite rightly that there are large scale, distributed teams using it as well).
Could probably even get away with RCS, but i'm not sure of too many projects that use that.
If you can find a business (with more than say, 3 employees) willing to use a 240 day trial as their enterprise virtualisation platform, you may have had an argument.
Slashdot Mirror
google "amd athlon windows 98"
Yeah, running on a quad. VTMB worked on Vista 64 on the same box though. WIll try your suggestion out though, cheers...
It matters, because games, especially as they get larger and more complex in the coming years (which is what 7 is designed for, rather than yesterday's software), will be dealing with larger data sets (memory management and file access), many more threads, to take advantage of multiple cpus (processor scheduling), etc.
A slow hard disk or fast hard disk will still see a performance improvement when used in a more intelligent manner. If you want to upgrade hardware, upgrade hardware. An OS is not a hardware replacement, and vice versa.
The idea of "levels" is retarded anyway. Fast disk i/o is needed for continuously streaming game worlds, which is where things are going.
Fair enough. The AMOUNT missing depends on hardware, granted. However, the fact that any ram is unusable at all, is due to the OS being 32 bit non PAE mode, which is a software issue. This was my point, i think we're referring to 2 different things...
DX is only part of the platform. DX doesn't cover stuff like file access, memory management, processor scheduling, etc...
Source? Doesn't match my experience, other people are reporting significant *improvements* in frame-rate when comparing XP and 7.
I've been running Vista 64 on an X-fi since 2007. I have zero issues. Ditto since I've upgraded to the RC.
Pfft. Until there's a decent RPG or flight sim, consoles have nothing for me.
If you have a single core box with less than 2gb, XP is probably as fast or faster.
If you have multiple cores, plentty of RAM (its CHEAP now, so if not why not), 7 will be quicker. Especially if you have a half recent 3d card, in which case much of the GUI is offloaded to it and its video memory...
Moving forward, the benefits of 7 over XP (or vista) will only become more apparent.
NWN2 barfs with a DirectX error (some missing string), No idea why VTMB dies (black screen, can only kill with ctrl+alt+del). Apparently the NWN2 problem is fixed in a build more recent than 7100, not sure about vampire.
But on the whole, if it works on Vista (and really, the jump from 98 to XP broke HEAPS more games than the jump from XP to Vista does), it will most likely work just fine on 7.
No, its a software/driver issue. Same box under a 64 bit or PAE O/S will see the full 4gb.
Not all drivers work with PAE (that were otherwise certified for desktop XP use). Hence it was not put into a consumer OS.
XP 32 bit can not use a full 4gb. It can only "see" 3.foo GB of ram, due to having only 32 bit address space, but needing to reserve some of that for purposes other than RAM. Eg, memory mapped IO.
Install 4GB ram into 3 different 32 bit XP boxes (hardware wise) and you'll likely have 3 different amounts of RAM available to the OS.
So you've tested it? I've been running the RC in a production environment on a few machines for a number of months now with no issues.
Dunno where you get yours, but the iridiums in my bike and turbo cage are $25au each.
So, its basically just like water injection then? People have been using water injection since at least WW2...
This is why we need to eliminate the oil, and build the engines using ceramics or teflon, or some other sort of friction-reduced/frictionless material.
If your plugin is only 32 bit, it will crap out in a 64 bit browser...
Anyone told the RIAA and MPAA (and their attorneys) yet?
Meanwhile, people have shit they want to get done (eg, streaming pirate hi-def video over their home network) and 802.11draft-N works just fine. Today. And has done for at least a year or two. I'm all for standards, but when it takes 3+ years for a standard to get a rubber stamp, when it works just fine already; something is wrong. From reading TFA, one of the hold ups mentioned was "acceptable co-existence language" mentioned in the draft. Sounds to me like the standards committee is broken.
Actually, you'll have more than that. Area of the circle = pi*r^2
As far as draft-N goes, my apple airport talks to a couple of Dell draft-N notebooks, a Thinkpad, my mac mini, and some china-spec draft-N card in my main home PC.
And in the meantime i've been able to actually stream high-def-ish content over 802.11n...
Back in the day, when I was about 15 and starting out trying to write demos (as in demoscene - we managed to replicate a few effects we'd seen back in the 90s from other big groups - after going "Wtf, how does that work?"), a friend and I started coding together.
The actual CODING is a one man job, sure; but when you're trying to solve a problem, having 2 minds work on the ALGORITHM gets you better results, quicker IMHO.
Sure, you'll possibly end up with the same algorithm/code eventually anyway, but having a second set of eyes and brain analyzing the process you're going to use picks up potential problems earlier. While you're thinking you're clever with some part of the problem, the other person is finding the glaringly obvious bug that you haven't seen (yet).
2c.
Could probably even get away with RCS, but i'm not sure of too many projects that use that.
If you can find a business (with more than say, 3 employees) willing to use a 240 day trial as their enterprise virtualisation platform, you may have had an argument.