Originally posted by: T2k
Originally posted by: TheJian
Originally posted by: T2k
Originally posted by: TheJian
Originally posted by: blackangst1
What am I missing? Isnt the Q6600 for not much more a better CPU overall?
Not at all. Every single benchmark will be won by the Q9300/X3320. Period. And that's not including the fact that it runs cooler, should overclock better based on the 45nm process (assuming a board that can hit 2000+fsb) AND comes with SSE4.1.
The faster FSB, 100mhz, and core improvements make it win 7% avg in everything.
http://www.xbitlabs.com/articl...y/core2quad-q9300.html
Q9300 vs. Q6600. Q6600 didn't win ONE test.
Look at the power consumption. Q9300 is 100mhz more. Now look at the overclocked ones at Q6600 3.6ghz vs Q9300 3.5ghz. Which you would probably need to get close to the Q9300. A full 70watts!
I might want the 3350 if I bothered though. Easier overclock without huge FSB.
But your wallet usually dictates what you do...LOL
Too bad you can't read your own link:
Although Yorkfield processors can overclock up to 4GHz (without any extreme cooling solutions involved), Core 2 Quad Q9300 cannot reach that frequency. Since the new quad-core generation started supporting 1333MHz bus, their multipliers got considerably lower. For example, Core 2 Quad Q9300 we have discussed today works with 7.5x multiplier, which doesn?t allow this processor to get past 3.4-3.5GHz because contemporary mainboards have pretty limited functionality when it comes to increasing the FSB frequency past 460-470MHz by quad-core CPUs. And this is actually even lower than the maximum frequency quad core processors from the Kentsfield family, including Core 2 Quad Q6600, can reach.
As a result, Core 2 Quad Q6600 may remain a better choice for overclocker systems, because it may run faster than Core 2 Quad Q9300 in some cases. Moreover, overclocking of previous-generation quad-core processors is a simpler procedure that doesn?t depend that much on the mainboard functionality.
So, it turns out pretty hard to make the final conclusion about the youngest quad-core Yorkfield processor. The new Core 2 Quad Q9300 is definitely a great product, but only until you get to overclocking. From the overclocking prospective we have to be more careful with our verdict and would call it an interesting but maybe not the most optimal choice.
So in other words contrary to your beliefs
unless you're running your CPU at stock speed actully Q9300 is the crappiest choice you can make out of all tested ones.
Too bad you assume one result means everyone has the same result. "pretty limited" does not mean "ALWAYS LIMITED".
http://www.ocforums.com/showth...552970&highlight=q9300
Barely in the wild and already 480fsb, nothing special. One search at ocforums...LOL. Something tells me you can do better with only one result hitting 480 and only at 1.33v with no mention of anything special going on with the chipset (with an added chipset cooler for my koolance I wonder what could be reached...or any other water/chipset). It's not like he froze stuff to get there. Do your homework before you buy your board. Or better, take my advice and buy the higher multi proc. Did you miss that recommendation? I'm thinking at 3.6ghz it's pretty tough for a Q6600 to beat it unless you like room heaters or like fires.
Emmmm, basic logic never been your forte, right?
480FSB still nothing but 3.6GHz - something that ANY Q66000 can do without any serious (chill, vapo, etc) cooling involved. Hey my primary work machine, a dwarfed dx5150 upgraded to Q6600 (w/ new miniATX mobo) at work runs at ~3.5GHz IIRC, just for the heck of it, with stock cooling.
If you have a better mobo with higher max FSB then a higher multiplier will make it EVEN BETTER. It's a fact, period. Weekend OC'ers and wannabes might think otherwise but in reality a 7.5x is a serious shortcoming and immediately excludes this CPU from any serious list.
The point was is it better than Q6600 (all things considered, SSE4, power consumption, performance), not is it the best overclocker out there. Once you consider SSE4 Q6600 is dead in my mind. That has the potential to make Q6600 look like a dual core.
The point is that
"better' it might be FOR YOU but hardly for anyone else because of the lower multiplier, the less cache etc. when you look at less [/b]
Once you consider SSE4 Q6600 is dead in my mind.
You're obviously some enthusiastic young fella, with little or no knowledge, I can see now.
That has the potential to make Q6600 look like a dual core.
By this idiotic 'comment I can tell you barely know anything about SSE2 vs SSE3 or even preliminary SSE3 vs SSE4 results. t
"By the way, we performed some additional tests and found out that the performance boost in DivX 6.7 depends a lot on the type of the encoded movie. In our test fragment featuring a battle scene Yorkfield processor was only 30% faster, however, the movie suggested by Intel, with water surface ripples being the major part of it revealed almost 70% performance boost in Yorkfield?s case."
This was in Oct 2007 (xbit Qx9650 article) for crying out loud. What happens 6 months from now? Adobe just release premiere SSE4.1 update. Wonder what will happen when people benchmark that?
http://www.anandtech.com/showdoc.aspx?i=3105
Nearly double performance with only a 200mhz speed advantage on a laptop (2.8ghz Penryn vs 2.6ghz merom). What's that equate to clock for clock? 75-80% faster? SSE4 will become potent. This was a sept 2007 article. Again, evidence of 2008 shaping up to be SSE4 aware and deadly to chips base on 65nm. Any more questions? :evil: :roll:
Ouch, for crying out loud -
you never actually read the articles you link???
Typical differences were sub-10% except few cases in
Intel-supplied benches where it peaked into 30-ish.
Here's what AT wrote about it
in January:
These SSE4-optimized situations are far more rare than the 1 - 8% increases we saw elsewhere, the point being that should application support develop, Penryn could do much better. Honestly though, we don?t expect a critical mass of SSE4 applications anytime soon, these sorts of things take a long time to materialize.
Ummm you just happened to
skip this part, huh?. Jesus, so pathetic.
Try reading the
facts instead of projecting your clueless wet dreams, seriously.
Watch the personal attacks. Consider this your only warning.
Markfw900 Anandtech Moderator.
Reply to Mod: OK, duly noted.
Good, why don't you edit and rephrase your comments before I have to.
Markfw900 Anandtech Moderator.
From the Xbitlabs.com Q9300 article:
"Overclocking, however, makes quad-core processors more ?power-hungry? than the dual-core ones. That is why if you intend to overclock, you should keep in mind that the thermal and power parameters of your CPU as well as the entire system will increase dramatically. However, the new Core 2 Quad Q9300 can still be considered highly economical even when running at the frequencies far beyond the maximum, compared with the overclocked Core 2 Quad Q6600 from the Kentsfield generation."
70watts or so at 3.6ghz. It goes higher as the mhz goes up. So if we're matching at 3750 (500fsb on 7.5 multiplier) you'll be worse than 70w likely. Once we can actually get these in volume (next month?) they are supposed to be price at $260 or so same as Q6600. So your cpu would cost a person a good $50 extra a year at around 8hr usage (a lot of us have our machine downloading all day, encoders/content creators would be bad here also, as that is what a quad is for isn't it? REAL WORK?). So over the life I get $150? 3yr's of your chip I can buy a 750GB today and count it free before my chip warranty runs out...LOL. Anandtech's QX9650 article agrees, 3ghz to 3ghz QX6800 "If you weren't dazzled by the performance improvements of Penryn, the reduction in power consumption is worth getting excited about."
http://anandtech.com/weblog/showpost.aspx?i=419
500 FSB a day later. That's up to 3750 now correct? Whats that take now a 3.9-4ghz Q6600 to match this?...Clock for clock they are NOT the same. These boards are not even optimized yet. Considering Duals reach 2100fsb, Anandtech just hit 2000fsb with quad, I'd say 2100fsb shortly is reasonable and I'm not even discussing the freak freezer/tweaker guys.
SSE4? Lets totally ignore that for a second shall we?
http://www.anandtech.com/cpuch...howdoc.aspx?i=3137&p=5
SSE2 is improved in DIVX by 31%. SSE2 last time I checked was WIDELY USED. "The performance difference is a little less significant here, with the
SSE4-less QX6850 taking about 31% more time to encode the input file than the QX9650.
Both of these are very real-world implementations of SSE4; unfortunately, it's tough to say how long it will be before we see widespread use of the new instructions."
Didn't say it would happen overnight. But nobody seems to be arguing that it WILL happen. But look at that SSE2 would you...ROFLMAO. 31% here TODAY!
SSE4 one paragraph before that? "On our QX9650, the full search with SSE4 enabled runs about 45% faster than with SSE2 only - impressive! Note also that the Penryn QX9650 offers better SSE2 performance in this test as well, coming in about 61% faster than the QX6850. The total performance increase from QX6850 SSE2 to QX9650 SSE4 in this test is an incredible 133%. Obviously, this is not going to be the norm in many other applications, but there's definitely some potential for meaningful optimizations in certain applications."
Was that 133% I saw there total? Don't forget the VERY REAL WORLD comment. As stated before SSE4 is now supported in Adobe Premiere CS3! Didn't these chips just hit the shelf? Already a major app jumping on? Already Divx jumping on? Virtual DUb? We're fighting over these things because they aren't even here in volume and already have support. I submit it won't be too long. Can you actually defend getting KILLED? Even if it takes a year? Do you think everyone upgrades yearly? How can the inevitable be ignored?
OH, same article overclocking page:"Our unlocked QX9650 had no problems hitting 333MHz x 12.0, for a final clock speed of 4.0GHz. We had to increase the stock voltage of 1.25V up to 1.40V to achieve it, but the overclock required no additional cooling beyond the standard Intel heatsink/fan. At lower voltages, 3.66GHz should be an easy target to reach.".
What kind of cooler do you need to hit 3.75ghz? It's NOT STOCK on Q6600. Q9300 scores another $70 savings (thermalright for you with fan?). You're chip is fast becoming SLOW AND EXPENSIVE isn't it?
Q9300 better for me but not anyone else? because of low multi (eh, 500fsb=3750mhz), because of low cache (xbit showed even 100mhz differnce = 8% victory for Q9300 over Q6600 and that with less cache...did you read the article?)? Even if I give you back the 100mhz it won't help you cover 8% (some benchmarks double digits!) in that article. I'll give you 3%, but that means I still win by 5% clock for clock with less cache.
Q6600 is at a dead end. Do I mean it sucks? Given the choice of Q9300 for a LOT of people yes. In the end considering the price of a heatsink to cool a Q6600@3750 or so ($50 best - $75 worst?), given the HUGE gaines in power (40w or so idle in that same anand article, 70w load so $50-70 a year for 3yrs another $150 totaled up), and SSE2 today (30%?) and SSE4 next year(? - 133% faster? Even if I take a quarter of that I'd take it no matter what...it's a no brainer all day) this Q6600 sucks. At a $200 savings roughly over 3yrs (heatsink+power savings) my chip would almost be free no? I'm not even counting performance outlined here. Still want to tell people to buy the Q6600? Q6600@3750 is a room heater. I don't want to be you in the summer.
I read ALL of the articles and apparently paid more attention than you did. I'm not much of an optimist either. More of a realist (clock for clock Q9300 is faster...SSE4 WILL be adopted, Q9300 IS cheaper to own over the life of the product). I'm not a young fella (ok, thanks, 37 is young now..LOL), and my knowledge is 8yrs as a pc business owner, and few thousand builds (20yrs of that), benchmarking beyond boredom etc.
Even your best case scenario Intels SSE4 garnered 1-8% running that same chip vs SSE2. Add that to the fact Q9300 is already faster by 5% clock for clock and Q9300 is 6-13% faster already. Not even considering the price how can you defend this choice? Heat? How can you defend it?
I don't dream. Wet or otherwise. I'm not going to waste any more of my time. Anyone who reads this should see the cost even at performance parity (which is NOT reality Q6600 loses. Period). We haven't even started the Q9xxx/X3xxx board optimizing for FSB (nvidia is starting hitting 500fsb) I expect more eventually. It just a bad choice to buy OLD stuff unless there is SOME benefit gained. I don't see one. NONE. FYI those tests above only show 2 (YEAH TWO) SSE4 instructions being optimized. There are 47 of them! Also note I DID say I'd buy the next step up X3350/Q9450. I guess you missed that so I'm saying it again. Look at Markfw900's post. How often do people hit 3750mhz on a Q6600? Both of these are budget chips, no? Why would someone want to pay more over the long run when debating a budget in their head? EVen at a $50 price hike for X3320's over Q6600 I've shown you'll more than make up for it after 1yr. That said I'd wait a month and pick this up at $250 instead of $290... Then it's a moot pricing issue.