Dude your channel is an absolute gold mine for quality nostalgic content. Please keep up the excellent work, your videos are highly entertaining! how much of a meme would it be if you did a vista era video…. do you know showing the OS and all its glory 😂
Edit: I've used FRAPS for benchmarking, not MSI Afterburner, because it didn't want to work properly. The Doom 3 performance really impressed me, and it seems too good to be true, but I've run the tests multiple times, having the same results. What are your thoughts on this?
The only trouble I have with this is how much would HT have affected the single core benchmarks. Quite a few P4s around this time had HT. Good video comparison.
I understand your concerns, but I've used a Pentium D, and disabled one core when testing single core performance. And yes, HT affects the performance in some games.
i had and used a p4 northwood with HT. it almost always made things run worse, or about the same. few games benefited from it, even ones (that i had anyway) with explict smp/smt support saw little if any improvement. overall multitasking generally did better. in the end i had it enabled more than disabled, because i frequently switched between games and a half dozen applications open at once.
Going the same way with Pentium III 1000 is enlightening...my expectations were exceeded as XP felt faster than it does on a Pentium 4 HT 3.2GHz with single channel DDR 333 (vs single channel 133MHz SDRAM) and with a Seagate U5 that was under half the speed of the laptop's HDD. Later through Cinebench 2003 I found that together they were legitimately nearly as fast as an Athlon XP 2600+ (and faster than a 2.2GHz Celeron overclocked to 3.1GHz)...in other words if a game supports SMP, those dual PIII can nearly match my Sempron 2800+. And while consuming only a bit more power (probably around 10W more at most). Apparently Quake III and *maybe* other games with the same engine support SMP (Doom 3's results are more than likely an example of that being idtech 4 - try Return to Castle Wolfenstein, Call of Duty as they are idtech 3), and in theory should work on early dual cores.
@@MidnightGeek99 I was lucky enough to pick the board (Abit VP6), 2GB SDRAM (sequential serials), CPU (also sequential) and heatsinks up for less than the board itself was going for on Ebay at the time - and from a small shop close to where I was going to college. Its one of things where I knew that I had to get it. Even if it made my budget tight for a fortnight.
I recently got a Thinkpad with a T7500 and a Intel X3100. I was debating if going 7 or 98SE since the gpu is so weak, and I discarded XP because I thought it wouldn't handle very well multicore, but I was wrong. I suppose XP would be the sensible choice, I never got problems with W32 games on XP and I still get the multi core support, plus I imagine I can change the core affinity to leave an entire core for the game. This was helpful, thanks.
I don't think that the video card has enough "juice" to use the T7500, but you can try :) I'm going to do a similar video, but this time using a Core 2 Quad, and testing from 1 core to 4 cores.
@@MidnightGeek99 It's very troublesome, because I have more or less enough CPU to run Half Life 2 decently but the X3100 can only run it in very low settings. That GPU is more suited to things like Return to Castle Wolfenstein, Max Payne and maybe Unreal 2, but it's better for being a Quake 3 and Half Life killer. A W98 would make more sense if it was a Pentium M instead a C2D, but it doesn't any kind of multicore support, hence the dilemma. I'm kicking myself right now because 15 years ago I got an HP laptop witth Pentium M and an ATI 9600, which would be perfect for a early-mid 2000 XP machine. And I sold it very cheap.
@@MidnightGeek99 Look forward to your next one - C2D v C2Q maybe? I have an E8400 @ 3.6 GHz with a HD4870 and an Audigy 2 ZS. It's such a fantastic machine. XP gaming at its best, for me! I do have a C2Q which I can run at the same speed in it but I much prefer the dual core, it just "feels" better.
@@ted-b Well, a C2Q will never offer a performance increase over a C2D in XP era games, but I can make this comparison in Windows Vista or 7, using newer games. Your build is awesome, it's not so advanced as to go way past the XP era, and it's powerful enough for playing any game from that period with all the bells and whistles. I'm also about the "feel", if a build does not feel right...I don't care about it's performance :)
@@MidnightGeek99 Yes, all the drivers are pretty much rock solid, that era of hardware suits XP so well it really is a joy to use. I have one of those 20" Dell monitors @ 1600 x 1200, the 4K of the day but with the "correct" aspect ratio, so the games look just right. Lovely! Anyway keep up the good work and keep having retro fun!
Remember that the Pentium D wasn't a true dual core. There's no shared L3 cache. Instead there's two L2 caches on each die which incurs penalties for core to core communication.
Well, it isn't a true dual core, it's 2 Pentium 4 glued together :) Also, you're right, the core to core communication was not great, Athlon 64 X2 was a real dual-core, and then the Core 2 Duo series from intel.
I've just did a test Athlon 64 2.2 GHz vs Athlon 64 X2 2.2 GHz (socket 939+AGP HD4650) with very similar results. In the vast majority of games, there is no difference, only one game where I saw little difference in exteriors (in interiors, there was no difference) was FarCry and even when FPS looked almost the same, game felt slightly better with less FPS drops. But generally, one superior single core CPU is better for these games than outdated DualCore platform like Pentium D or Athlon X2 for socket 939. When it comes to Core2 Duo, situation is very different, that architecture is superior, so you see a massive difference, but not because of more cores, but because of much better single core performance. I was very disapointed by performance of sc 939 dual core and by that even 2007 games still can't really use 2 cores. Another weird thing I noticed is that 2005 FEAR runs very badly on everything, but I remember playing it on some Pentium 4 with some 7600GS or something, fortunately, I probably didn't know fraps in that time yet, so I don't know what FPS was there, but it's weird that it doesn't work well neither on dual core and 2008 GPU, you can clearly see that switch to Core 2 Duo and PCI-e card was and Upgrade with big U, you have like 4 times better performance. 😀 I haven't tested Doom 3, but I remember that in Quake 4 (same engine) there is some dual core settings in game.
I strongly believe that you need to re-do the test's but this time with a standalone Pentium 4 processor. I did a few benchmarks by myself with the Pentium D 945 with 1 core disabled and a Pentium 4 clocked at 3.4 ghz, the standalone P4 crushed the pentium D performance...
Also, for this test to be valid when using 2 CPUs, one single core and one dual core, the single core one need to have the same performance as the dual core with just one core enabled, otherwise you won't see a correct difference in FPS values.
I don't have a P4 3.2 right now, but it will be strange for the P4 to be faster than the Pentium D 9xx, with one core disabled, in theory they should have the same performance.
@Lurch7861 Maybe I'll do some tests, a P4 vs Pentium D with just one core. I agree with the P3 vs P4, but not 100%, the Northwood period was quite OK for intel...compared to what AMD had back then. Core 2...yeah, it's way...everything, of course, but we start getting into Windows Vista territory there.
Crazy. I had an 8600GTS back in the day (a bit worse than the X1900XTX used here, if recall) and I remember F.E.A.R. running FAR better on my Athlon 64 X2 4200+, and that thing was only 2.2GHz. It really is insane just how bad Netburst was.
A64 X2 4200+ compared to this 940 is way faster, but no worries, I'll redo the tests using a Core2 and a better GPU, to have clearer results...and maybe add other games also.
i suggest doing the test with as many the same parts as possible, or as close as possible if you want to know with less/no doubt the cpu was the bottleneck.
Yes, and GeForce were better in Doom 3, that's how things were back then. Althouth the Radeon X850 XT was able to match the 6800 Ultra, once ATI drivers became more mature.
Dude your channel is an absolute gold mine for quality nostalgic content. Please keep up the excellent work, your videos are highly entertaining! how much of a meme would it be if you did a vista era video…. do you know showing the OS and all its glory 😂
Thanks! Well, one of my next videos is going to be about Vista...so, meme incoming I guess :)
Edit: I've used FRAPS for benchmarking, not MSI Afterburner, because it didn't want to work properly.
The Doom 3 performance really impressed me, and it seems too good to be true, but I've run the tests multiple times, having the same results.
What are your thoughts on this?
Excellent video
Thanks! It's nice hearing from you again :)
Wow man, great video. Keep it up bud!
Earned my sub.
Thank you :)
The only trouble I have with this is how much would HT have affected the single core benchmarks. Quite a few P4s around this time had HT. Good video comparison.
I understand your concerns, but I've used a Pentium D, and disabled one core when testing single core performance.
And yes, HT affects the performance in some games.
@Lurch7861 Yeah, but the Pentium D does not have HT...it would have been nice to have HT though.
i had and used a p4 northwood with HT. it almost always made things run worse, or about the same. few games benefited from it, even ones (that i had anyway) with explict smp/smt support saw little if any improvement. overall multitasking generally did better. in the end i had it enabled more than disabled, because i frequently switched between games and a half dozen applications open at once.
Going the same way with Pentium III 1000 is enlightening...my expectations were exceeded as XP felt faster than it does on a Pentium 4 HT 3.2GHz with single channel DDR 333 (vs single channel 133MHz SDRAM) and with a Seagate U5 that was under half the speed of the laptop's HDD.
Later through Cinebench 2003 I found that together they were legitimately nearly as fast as an Athlon XP 2600+ (and faster than a 2.2GHz Celeron overclocked to 3.1GHz)...in other words if a game supports SMP, those dual PIII can nearly match my Sempron 2800+. And while consuming only a bit more power (probably around 10W more at most).
Apparently Quake III and *maybe* other games with the same engine support SMP (Doom 3's results are more than likely an example of that being idtech 4 - try Return to Castle Wolfenstein, Call of Duty as they are idtech 3), and in theory should work on early dual cores.
Dual Pentium 3s....wet dream :) I'm going to do similar tests, but with faster CPU and GPU.
@@MidnightGeek99 I was lucky enough to pick the board (Abit VP6), 2GB SDRAM (sequential serials), CPU (also sequential) and heatsinks up for less than the board itself was going for on Ebay at the time - and from a small shop close to where I was going to college.
Its one of things where I knew that I had to get it. Even if it made my budget tight for a fortnight.
Great channel; you're a natural
Thank you, have fun :)
I recently got a Thinkpad with a T7500 and a Intel X3100. I was debating if going 7 or 98SE since the gpu is so weak, and I discarded XP because I thought it wouldn't handle very well multicore, but I was wrong. I suppose XP would be the sensible choice, I never got problems with W32 games on XP and I still get the multi core support, plus I imagine I can change the core affinity to leave an entire core for the game. This was helpful, thanks.
I don't think that the video card has enough "juice" to use the T7500, but you can try :)
I'm going to do a similar video, but this time using a Core 2 Quad, and testing from 1 core to 4 cores.
@@MidnightGeek99 It's very troublesome, because I have more or less enough CPU to run Half Life 2 decently but the X3100 can only run it in very low settings. That GPU is more suited to things like Return to Castle Wolfenstein, Max Payne and maybe Unreal 2, but it's better for being a Quake 3 and Half Life killer. A W98 would make more sense if it was a Pentium M instead a C2D, but it doesn't any kind of multicore support, hence the dilemma.
I'm kicking myself right now because 15 years ago I got an HP laptop witth Pentium M and an ATI 9600, which would be perfect for a early-mid 2000 XP machine. And I sold it very cheap.
A fun experiment, nice one! 😃
Thank you, I was really excited about this, because it loomed in my head for a while now :)
@@MidnightGeek99 Look forward to your next one - C2D v C2Q maybe? I have an E8400 @ 3.6 GHz with a HD4870 and an Audigy 2 ZS. It's such a fantastic machine. XP gaming at its best, for me! I do have a C2Q which I can run at the same speed in it but I much prefer the dual core, it just "feels" better.
@@ted-b Well, a C2Q will never offer a performance increase over a C2D in XP era games, but I can make this comparison in Windows Vista or 7, using newer games.
Your build is awesome, it's not so advanced as to go way past the XP era, and it's powerful enough for playing any game from that period with all the bells and whistles.
I'm also about the "feel", if a build does not feel right...I don't care about it's performance :)
@@MidnightGeek99 Yes, all the drivers are pretty much rock solid, that era of hardware suits XP so well it really is a joy to use. I have one of those 20" Dell monitors @ 1600 x 1200, the 4K of the day but with the "correct" aspect ratio, so the games look just right. Lovely! Anyway keep up the good work and keep having retro fun!
@@ted-b Thank you. What model is your monitor? I also have a 4:3 display @ 1600x1200, a 21" FlexScan S2100, it's awesome.
Remember that the Pentium D wasn't a true dual core. There's no shared L3 cache. Instead there's two L2 caches on each die which incurs penalties for core to core communication.
Well, it isn't a true dual core, it's 2 Pentium 4 glued together :)
Also, you're right, the core to core communication was not great, Athlon 64 X2 was a real dual-core, and then the Core 2 Duo series from intel.
I've just did a test Athlon 64 2.2 GHz vs Athlon 64 X2 2.2 GHz (socket 939+AGP HD4650) with very similar results.
In the vast majority of games, there is no difference, only one game where I saw little difference in exteriors (in interiors, there was no difference) was FarCry and even when FPS looked almost the same, game felt slightly better with less FPS drops. But generally, one superior single core CPU is better for these games than outdated DualCore platform like Pentium D or Athlon X2 for socket 939. When it comes to Core2 Duo, situation is very different, that architecture is superior, so you see a massive difference, but not because of more cores, but because of much better single core performance.
I was very disapointed by performance of sc 939 dual core and by that even 2007 games still can't really use 2 cores. Another weird thing I noticed is that 2005 FEAR runs very badly on everything, but I remember playing it on some Pentium 4 with some 7600GS or something, fortunately, I probably didn't know fraps in that time yet, so I don't know what FPS was there, but it's weird that it doesn't work well neither on dual core and 2008 GPU, you can clearly see that switch to Core 2 Duo and PCI-e card was and Upgrade with big U, you have like 4 times better performance. 😀
I haven't tested Doom 3, but I remember that in Quake 4 (same engine) there is some dual core settings in game.
I will do another video, this time with a Core 2 Duo :)
I strongly believe that you need to re-do the test's but this time with a standalone Pentium 4 processor. I did a few benchmarks by myself with the Pentium D 945 with 1 core disabled and a Pentium 4 clocked at 3.4 ghz, the standalone P4 crushed the pentium D performance...
Hmmm...have you then tested the P4 against the Pentium D 945 with both cores?
Also, for this test to be valid when using 2 CPUs, one single core and one dual core, the single core one need to have the same performance as the dual core with just one core enabled, otherwise you won't see a correct difference in FPS values.
I don't have a P4 3.2 right now, but it will be strange for the P4 to be faster than the Pentium D 9xx, with one core disabled, in theory they should have the same performance.
@Lurch7861 Maybe I'll do some tests, a P4 vs Pentium D with just one core.
I agree with the P3 vs P4, but not 100%, the Northwood period was quite OK for intel...compared to what AMD had back then.
Core 2...yeah, it's way...everything, of course, but we start getting into Windows Vista territory there.
Crazy. I had an 8600GTS back in the day (a bit worse than the X1900XTX used here, if recall) and I remember F.E.A.R. running FAR better on my Athlon 64 X2 4200+, and that thing was only 2.2GHz. It really is insane just how bad Netburst was.
A64 X2 4200+ compared to this 940 is way faster, but no worries, I'll redo the tests using a Core2 and a better GPU, to have clearer results...and maybe add other games also.
Back then , I thought Pentium D dual core mean double performance compare to Pentium 4 at same clock speed . For me , It was a dreaming cpu .
:)) Double damage, or Double trouble.
i suggest doing the test with as many the same parts as possible, or as close as possible if you want to know with less/no doubt the cpu was the bottleneck.
I've done all the tests using the same components :)
I just installed XP on a HP Z400 3.3ghz on a xeon quad core basically a I7
Do you have a dedicated video card with it?
2:17 which game is this?
Half-Life 2
ATI gpu's are way better and better optimized for Farcry one :)
Yes, and GeForce were better in Doom 3, that's how things were back then.
Althouth the Radeon X850 XT was able to match the 6800 Ultra, once ATI drivers became more mature.
@@MidnightGeek99 no its really consistent with most if @ ALL ati gpu's that can run this game vs nvidia