Wow, Price vs Perf right now in PC's is nice...

Albuquerque said:
What are the PCI and southbridge going to do with memory? DMA transfers, maybe? The reality is it's not going to be doing DMA transfers at the exact same time that the CPU is 100% utilizing the memory bus, so it's still a 100% non-issue. And DMA is the only point where the CPU doesn't intervene in memory access between peripheral devices and main memory.
Click to expand...

And I thought we all saw how core 2 is far from 100% utilizing the memory bus. You were talking about FSB cap so far. Anyway that is why DMA priority modes and bus mastering exist.

Also I am not sure about power draw, since different memory speeds will most likely have different chips.
 
Last edited by a moderator:
Putas said:
And I thought we all saw how core 2 is far from 100% utilizing the memory bus. You were talking about FSB cap so far. Anyway that is why DMA priority modes and bus mastering exist.
Click to expand...

Yes, so? You keep arguing exactly why we SHOULDN'T get faster memory. You are already in agreement that C2 doesn't need it, that it will almost never fully saturate it, and how DMA transfers are already going to have plenty of room to operate.

So why do you keep telling me it's a good idea again?

And how does this correlate to the subject of computers being a great price versus the performance they deliver in today's market?
 
Albuquerque said:
And it isn't my rig, so I'm not buying jack. And I'm not suggesting he buy anything different, either.
Click to expand...


Good for you, and props to you for standing firm against a blizzard of typical Internet nonsense regarding PC components. :yep2: Sorry I don't have anything more substantive to add to the thread but I like to see sanity prevail once in a while.
 
Putas said:
Benchmarks show limited gains because core2 in common configurations is not bandwith starved. Lower memory clocks below 666 means same limited performance drops. Still it is scaling slowly but steadily with memory up to 2 effective GHz (faster was not yet tested).
Click to expand...
Actually, that's not true. Lowering the memory clock below the FSB shows much, much bigger performance drops.

There's some great data at Anandtech. With the 1066MHz FSB, you'd see a huge drop in performance going down to to DDR2-400 mem, but get barely any boost going up to DDR2-1066.
 
Mintmaster said:
Actually, that's not true. Lowering the memory clock below the FSB shows much, much bigger performance drops.

There's some great data at Anandtech. With the 1066MHz FSB, you'd see a huge drop in performance going down to to DDR2-400 mem, but get barely any boost going up to DDR2-1066.
Click to expand...

After more reading I see I was wrong:oops:. Still I would correct you - in real applications with 1066 FSB and DDR2 533 going to 400 or 1066 makes aproximately equal diffrence. And with such low speed maybe the cpu is finally starving?
 
Sorry to interupt the fist fight but I have to butt in and say, I love the price vs performance ratio of today's PCs. I'm incredibly happy with my Q6600 @ 3GHz (stock @ 2.4GHz). :p

PS: I have DDR2 1066MHz memory and it's very good at overclocking.
 
Price:performance is great right now in the market. I did opt for DDR2 800 for the reason I got 2GB of the stuff for under $40--name brand at that (OCZ Platinum Revision 2 with tight timings).

The upgrade path that current PCs offer are nice as well. Pick up a C2D today, and in 12-24 months you could drop in a quad core (non-Nehalem) with a noticable frequency increase, more cache, and more cores for a modest sum.
 
Putas said:
If you will review the thread you will see I never did. It all just happend in your fantasy.
Click to expand...

Quite the opposite. Every reply of yours in this thread has been why it's "good" to get memory faster than the FSB. Which we've now shown to be clearly and entirely false. Perhaps you should pay more attention?

Back to the OP, I'm going to agree with Luna about upgradability -- and I normally wouldn't. Someone who came into the C2D platform at the very beginning has some fantastic upgrade opportunities; I don't believe that's happened for an Intel person since the 486 era.

I'm quite sure it's all because of competition; unfortunately I see the competition dwindling now. I fear that the next 6 months will be the end of this plateau, and we'll be going back to incremental increases for monumental price changes. Hopefully I'm wrong! :cry:
 
I was thinking earlier today, I can buy 4GB of memory now at HALF the price I bought 2GB last year at around the same time frame. I think its more than half actually. Crazy how fast memory prices have fallen. It's a great thing though, I can actually think about going to 4GB now without any real concern about price since its slow low.

I think CPU prices still need to come down a bit for me personally, mainly in the quad core market. My E6300 has OC'd extremely well and therefore none of the current dual cores offer much in the way in improvement other than 4MB instead of 2MB of L2 cache, and a new name.

Video cards are looking a lot better with the arrival of the HD 3870 and that forcing Nvidia to give us the 8800GT 512Mb and 256MB cards.

Sadly I don't think this is going to last much longer. I think soon we might see price increases in all these markets actually.
 
Albuquerque said:
Back to the OP, I'm going to agree with Luna about upgradability -- and I normally wouldn't. Someone who came into the C2D platform at the very beginning has some fantastic upgrade opportunities; I don't believe that's happened for an Intel person since the 486 era.
Click to expand...

I said the same things elsewhere :p

Typically, by the time we see processors with 70-100%+ performance increase, the range I would consider a reasonable one to upgrade for in regards to performance issues, you have to get a new MB... and often RAM, and maybe even a GPU and PSU.

Right now with the C2D you can slap in a cheap Dualcore with 4MB of memory, and when performance becomes an issue you can pick up an LGA775 C2Q with 12MB L2 cache and frequencies in the 3.6GHz+ range. Depending on the C2D you pick up (or previously picked up) you could be looking at a serial performance gain well over 70%+, plus 2 additional CPU cores.

Even now, if I get an E6750 (2 cores, 4MB cache, 2.66GHz, 65nm) it is likely in the next 24 months to be able to find a solid C2Q (4 cores, 12MB cache, 3.6GHz+, 45nm) that overclocks well. While you may not break the 70% increase in single threaded performance, with multi-threaded games and applications you may still come out quite ahead in such apps.

Sockets change so fast that CPU upgrades are a crap shoot (Nehalem will be ushering in new sockets in 2008 for the integrated memory controller), but right now LGA775 does seem to offer some hope that a C2D investment now or recently will have some legs to remain relatively competitive in the market. That doesn't happen very often. I hope, but doubt, the trend will continue.
 
i always buy faster ram, but thats so i can lower the muti and up the FSB/RAM at a 1:1 ratio (assuming i keep around the same overall clocks)
 
ShaidarHaran said:
Who cares?

Real world tests (like those seen at Tech Report) show fully-loaded similarly-configured single-GPU systems drawing ~300W. PSU calculators are a joke.
Click to expand...

300W actual or apparent power? If 300W actual power, well PSU are rated in apparent power. Depending on their efficiency, they may get nowhere near their rated power. I think about 80% efficiency is about typical under high loads, so that 500W power supply is now down to 400W. (cheaper power supplies have worse efficiencies as well, and I think generally the efficiencies under low utilizations are worse...age also effects efficiency)
 
Also running PSU at near maximum load will not do it any good, not to mention that at higher loads its efficiency drops leading to even higher power use, heat generation and noise.
 
Fox5 said:
300W actual or apparent power? If 300W actual power, well PSU are rated in apparent power. Depending on their efficiency, they may get nowhere near their rated power. I think about 80% efficiency is about typical under high loads, so that 500W power supply is now down to 400W. (cheaper power supplies have worse efficiencies as well, and I think generally the efficiencies under low utilizations are worse...age also effects efficiency)
Click to expand...

~300W from the wall; It's not hard to find these reviews. Which is why the 50% bigger PSU I specced for the boss isn't going to puke and die.

Further more, better quality PSU's will also have Active Power Factor Correction, which further increases the efficiency at higher load. And if you want to argue about how PSU's that are specced within 80% of the machine's current draw are going to die, then tell it to all the OEM's out there who do the same (and have done the same for decades) for the equipment they sell with 3 year and 5 year warranties. Like Dell, IBM, HP/Compaq, Asus, you name it.

There's a 275W PSU in my Dell Optiplex GX620, and it's running a 2.8Ghz Pentium-D, 4GB of ram, a 7200RPM drive and an x600. It's been running on my office desk for two years, and I can 100% guarantee that a good six months of that was at full load. The Optiplex 745 right next to it has the same exact PSU, but with a 3.4Ghz Pentium-D, 4GB of ram, a 7200 RPM drive and a DVDRW burner. For it's year of life on my desk, it has probably seen six months worth of 100% usage too.

And these are power supplies that are like 2" across, by about 2" deep, by about 7" long: slimline platform. And I don't ever expect any problems out of these boxes, because our entire company has probably more than a thousand of these units with zero PSU issues.

So in other words, research would be your friend right about now.
 
Fox5 said:
300W actual or apparent power? If 300W actual power, well PSU are rated in apparent power. Depending on their efficiency, they may get nowhere near their rated power. I think about 80% efficiency is about typical under high loads, so that 500W power supply is now down to 400W. (cheaper power supplies have worse efficiencies as well, and I think generally the efficiencies under low utilizations are worse...age also effects efficiency)
Click to expand...

It's the other way round, isn't it? 80% efficiency means that it will deliver 500W but draw 625W from the socket (under full load, that is). Anything else would make no sense whatsoever.
 
Fox5 said:
300W actual or apparent power? If 300W actual power, well PSU are rated in apparent power. Depending on their efficiency, they may get nowhere near their rated power. I think about 80% efficiency is about typical under high loads, so that 500W power supply is now down to 400W. (cheaper power supplies have worse efficiencies as well, and I think generally the efficiencies under low utilizations are worse...age also effects efficiency)
Click to expand...

As L233 said, efficiency doesn't reduce a 500W PSU to 400W, it means it pulls more power from the wall than the load placed upon it.
 
ShaidarHaran said:
As L233 said, efficiency doesn't reduce a 500W PSU to 400W, it means it pulls more power from the wall than the load placed upon it.
Click to expand...

Whether that happens because it outputs 500W but draws 600W or because it draws 500W but outputs 400W, the statement that "it pulls more power from the wall than the load place upon it" is still true. (just in the latter case it'll fail before the first case) Your statement alone isn't a contradiction to what I said, unless you know for a fact that the industry rates their power supplies off of output power and not power drawn from the wall. AFAIK in industrial applications, the ratings are always power drawn from the wall.
 
Fox5 said:
Whether that happens because it outputs 500W but draws 600W or because it draws 500W but outputs 400W, the statement that "it pulls more power from the wall than the load place upon it" is still true. (just in the latter case it'll fail before the first case) Your statement alone isn't a contradiction to what I said, unless you know for a fact that the industry rates their power supplies off of output power and not power drawn from the wall. AFAIK in industrial applications, the ratings are always power drawn from the wall.
Click to expand...

The way you've phrased it makes it sound as though a 500W PSU is only capable of delivering 400W, which is simply not true.
 
You must log in or register to reply here.
Back
Top