E7300 - Where does this chip fit?

[TheJian]

I'm wondering because I've been looking around and coming up with just about nothing.
I show the E4700 in a cpu compatibility chart for the Gigabyte EP35-DS3R.
http://www.gigabyte-usa.com/Su...el.aspx?ProductID=2637

Since Newegg has the e4600 price at $140 and E4500 at $125 a good guess puts the E4700 at above $155. Since as you go up the spread further up the chain. The E4700 is 2.6ghz with 2mb cache. E7300 (if we're to believe an ES post available out there) runs 2.66 3mb 1066fsb. We have NO E4700 pricing or availability. Wouldn't this chip be seriously overlapping E4700 if the E7300 is a may chip? We're almost in April and no E4700 (but it's supported in bios, no mention of E7xxx anything nor is a E5xxx chip shown).

http://www.techfuzz.com/roadmaps/2008.aspx
This place shows E5xxx series with 3MB cache 1066fsb. Also has E7200 down 2nd half (that's July or later so it's doubtful we'll see E7300 before this). Since none of the E4xxx series has VT I'm wondering if the true replacements for E4xxx series is this 5xxx series without VT and the E7xxx series has VT (which would put it more expensive and would make more sense making it perhaps replace the E6xxx series which has VT also). Note the APRIL post in that link showing E5xxx series replacing E4xxx. Note they have 3mb and 1066fsb. Obviously E7300 isn't replacing E5xxx correct?

I tried to find other roadmaps confirming or denying, anyone else have something on a better roadmap (official?)? IF this place is to be believed the E7300 won't be cheap as chips as some in here have speculated. It will be E6xxx priced not E4xxx. Of course that just means you'll all be looking for E5xxx chips instead...

 

[harpoon84]

Once the E4700 comes out, it will take the pricepoint of the E4600, and the E4600 will take the pricepoint of the E4500.

The E7300 (or E7200) could very well overlap with the E4700, just as E8200/8400 pricing overlaps with E6550 and E6750 pricing (of course the E8xxx chips are price gouged to hell).

So by May or June we should be looking at (in 1000qty prices):

E4600 - $113
E4700 - $133
E7200 or E7300 - $133
E8200 - $163
E8400 - $183
E8500 - $266

I actually think the E5xxx and E7xxx you mentioned are the same thing. Obviously Intel has decided that E7xxx sounds 'better' than E5xxx (bigger is better remember ). Actually it would make sense since the E7xxx chips should be slightly faster than the E6xxx chips clock for clock.

 

[TheJian]

It wouldn't make sense to you that the E5xxx series would replace E4xxx because of NO VT? I would think the E7xxx series would have VT like the E6xxx thereby making it more expensive. You got any link showing E7xxx coming before Half2 this year?

I see the QX9770 in the cpu list for the bios also. This chip can't be bought and is a ways away. After all the QX9650 just came out and the QX9770 was just added in the latest bios which is a beta I guess supporting FSB1600 chips. It seems to me if these E7xxx chips were around the block motherboard makers would be adding them to the same bios as the beta bios for QX9770. How long ago did the E3110 come out in retail? It was supported in a bios from DEC 2007. While the QX9770 beta bios came out 2/5/08. It would seem to me if these were coming in may they'd be in the beta bios.

Any link showing E7xxx replacing E4xxx instead of E6xxx chips or is that just an assumption? I've seen people saying this but nobody showing where they got this thought from

 

[harpoon84]

http://vr-zone.com/articles/Intel_Desktop_CPUs_Price_Cut_Schedule/5510.html

No mention of E5xxx.

The E7xxx is the E4xxx replacement. Just look at the price. E6xxx chips will be phased out once E8xxx production ramps up.

 

[TheJian]

Again compiled like the site I pointed to. But nonetheless does verify my thinking that it's over $133. That's a E7200 in that list for Q2 not an E7300 with 10x multi. Also Q2 could be june for all we know.

Having said that, I wonder which roadmap page is right. Why would the other site show both a 5xxx model and e7200 model? Either way all of these chips will be great for bang/buck purchasers. Could be the same but weird showing both.

Still glad I just bought another E3110 (for dad this time) Unfortunately it was $228 (incl shipping my other was $209) at excalibur. Then again they only had 107 and I'm not sure I REALLY got it yet. My other shipped but this one was over the weekend. That 107 doesn't show how many got ordered. They won't be pulled until today so I could be #1000 if tons of people found out about them Friday night. :Q I already told my dad I bought too...I might have to tell him he's still getting my old E4300 instead...LOL He'll probably cry this time. I broke his heart the first time when ewiz ran out on me, just hours later. I took too long looking at benchmarks trying to justify it

 

[TheJian]

Just noticed that chip E7200 at VR shows 800fsb. The other site shows 1066 for E7300 and so does the ES sample pics out there on the wibbly. Must be a mistake. Or that's a funky multiplier...

 

[harpoon84]

It's a typo. Obviously you can't get 2.53GHz from an 800FSB.

Wheter its the E7200 or E7300 that is launched, does it really matter? 2.53GHz vs 2.66GHz, big freaking deal, its still a 45nm Wolfdale 3-M chip. The VR-Zone article is from January, Intel could very well have changed their plans since then, they may even launch the E7300 and E7200 concurrently, or perhaps the E7300 will be launched at a later date, who knows.

 

[TheJian]

Originally posted by: harpoon84
It's a typo. Obviously you can't get 2.53GHz from an 800FSB.

Wheter its the E7200 or E7300 that is launched, does it really matter? 2.53GHz vs 2.66GHz, big freaking deal, its still a 45nm Wolfdale 3-M chip. The VR-Zone article is from January, Intel could very well have changed their plans since then, they may even launch the E7300 and E7200 concurrently, or perhaps the E7300 will be launched at a later date, who knows.

Uhh...It's a BIG freaking deal to anyone without a board that supports 1/2 multipliers. That was my point. As a reseller your customer would crap a brick if you sold them an incompatible chip with their shiny new board...LOL. Maybe everyone will do it with a bios, but still people should at least be aware of it and check first rather than causing fry's to get more white labels all over their products... This will also be a performance hit probably (I'm too tired to do math). Another reason I like any chip that works with a good speed at perfect clocks. Like a chip that runs where you want at 1600fsb. Perfect for DDR2 800. Of course you can do the math and change this to keep them perfect but that's just one easy example for people. A few % here, a few there, it all adds up. Whether you're screwing yourself or setting yourself up for running on all cylinders.

It also means the E7300 won't be $133. I'll go out on a limb and say both will cost more than $160 after a few days on the market. Unless you are believing these won't be incredibly popular. Newegg's price on the E8400 is now $260 (up another $10 still not in stock). Popularity = pain. I'm thinking they'll be even more popular than E8400 due to supposed price. Also with a lot of E6750/6850's out there still I can't see Intel letting many E8400/8500's loose or a lot of people having shelf stock of the old ones will freak. The same could happen and cause a delay with E7xxx series. You'd probably want 4xxx/6xxx gone considering both are moot and not desired with .45's cooler and more overclockable.

 

[harpoon84]

Can you list which mobos don't support 1/2 multipliers? All P35/X38/780i (basically all compatible Intel chipsets sold today) can run 1/2 multipliers. I'm not sure if older P965 chipsets can support them, my Asus P5B Deluxe can run 45nm chips so I assume with a BIOS update older chipsets can run 1/2 multis as well.

I'm not sure what performance hit you're talking about either, or 'good speed at perfect clocks' whatever that means. Sorry, you're gonna have to explain that bit again. :/

As for the price for the chip, why are you so obsessed about it? Look, if the E7xxx series will indeed replace the E4xxx series, then the top SKU will be $133, the next one down will be $113. This is how it simply works. We still don't know if the E7300 or E7200 will end up being the top SKU. If high popularity means it costs more, too bad, I'm sure after a while supply and demand will balance out.

 

[dasracht]

780i boards don't support .5 multipliers? I never knew that. So does that mean there would be no difference between the q9450 (8x) and the q9550 (8.5x) on my 780i?

Sorry to go off topic, but an answer would help me out a ton.

 

[Denithor]

Originally posted by: dasracht
780i boards don't support .5 multipliers? I never knew that. So does that mean there would be no difference between the q9450 (8x) and the q9550 (8.5x) on my 780i?

I think you misread the post...

Originally posted by: harpoon84
Can you list which mobos don't support 1/2 multipliers? All P35/X38/780i (basically all compatible Intel chipsets sold today) can run 1/2 multipliers.

Now, I have an issue with this statement:

Originally posted by: harpoon84
Actually it would make sense since the E7xxx chips should be slightly faster than the E6xxx chips clock for clock.

Somehow I don't think so, in the vast majority of applications. The 4MB cache on the E6xxx series should make up for any slight speed advantage from the Penryn architecture.

Now, if overclocking is taken into account...yeah, probably the E7xxx chips will be faster (but only due to higher max clockspeed, not work/cycle advantages).

 

[dasracht]

ah yeah I did misread, thanks for the correction!

 

[harpoon84]

Originally posted by: Denithor
Now, I have an issue with this statement:

Originally posted by: harpoon84
Actually it would make sense since the E7xxx chips should be slightly faster than the E6xxx chips clock for clock.

Somehow I don't think so, in the vast majority of applications. The 4MB cache on the E6xxx series should make up for any slight speed advantage from the Penryn architecture.

Now, if overclocking is taken into account...yeah, probably the E7xxx chips will be faster (but only due to higher max clockspeed, not work/cycle advantages).

Cache is overrated, especially above 2MB, which is the point of diminishing returns on the Core 2 architecture. The average difference between 4MB and 2MB of L2 cache is only3.5% according to Anandtech. Obviously the difference between 4MB and 3MB would be even smaller, under 2%.

I am basing my assumption that E7xxx will be faster than E6xxx chips based on the findings from this article: http://www.xbitlabs.com/articl...uad-q9300_8.html#sect1

Consider that a Q9300 (essentially a QC version of the Wolfdale-3M) when overclocked to 3.5GHz generally matches or exceeds the performance of a 3.6GHz Q6600. Also consider that at stock 2.5GHz, it is 7% faster than a 2.4GHz Q6600 despite only being clocked 4.1% higher. If a Q9300 can outperform a Q6600 per clock, despite a smaller cache, then clearly this is due to the architectural improvements from the Penryn shrink.

 

[TheJian]

Originally posted by: harpoon84
Can you list which mobos don't support 1/2 multipliers? All P35/X38/780i (basically all compatible Intel chipsets sold today) can run 1/2 multipliers. I'm not sure if older P965 chipsets can support them, my Asus P5B Deluxe can run 45nm chips so I assume with a BIOS update older chipsets can run 1/2 multis as well.

I'm not sure what performance hit you're talking about either, or 'good speed at perfect clocks' whatever that means. Sorry, you're gonna have to explain that bit again. :/

As for the price for the chip, why are you so obsessed about it? Look, if the E7xxx series will indeed replace the E4xxx series, then the top SKU will be $133, the next one down will be $113. This is how it simply works. We still don't know if the E7300 or E7200 will end up being the top SKU. If high popularity means it costs more, too bad, I'm sure after a while supply and demand will balance out.

With regards to memory I'll say google it I'm tired. try something like "running memory sync fsb best 1:1" ...That should narrow a search I'd think. While there are cases where it may not be true it's a good rule to follow (800 goes into 1600 nice and even, nobody will wait...That's as simple as my tired brain can say it, another 900/1800 etc). Never say never..but...

With regards to boards. I'm thinking I can't tell you what the board makers haven't even proven yet (1/2 multi chips just aren't out yet, how could I predict that?). I did say board bios updates would maybe solve this (your VRM's a different issue I guess). I can't think of the last time Intel ran a 1/2 multi so I'll reserve the answer to that until I see what motherboard makers do. Suffice to say people should check their boards website before buying a chip. If the chip ain't on your list don't buy and expect it to work. Penryn has already bit a few people (you can google more I guess). I just remember this maximum pc one found looking up someones evga. It was an easy find. http://www.maximumpc.com/artic...s_broke_penryn_support
Oh and yeah, you should have the best of luck with Intel based boards. After all they can call the chipset dept and say hey we have 1/2 multi's coming soon, support it. That comment should have been directly aimed at Nvidia people. My bad.

Umm price? Why are you so upset by me telling you it will be higher than Intels MSRP for a while? Reality really piss you off that much? Everything comes down with time. :Q No argument there. That wasn't the point of any price related post from me. The point was bang for buck goes down if we can't get it at or near Intel SKU price. I mean at $260+ for E8400 I start thinking about a quad. Are you going to start quoting me Intel's $183 price for it now... If you're one of the few who doesn't have to obsess over price congrats, you have more money than most of us I guess. Oops, pricewatch now shows only 3 with e8400 $310 lowest. Is that right? If people still want a e3110 excalibur had 107 Friday for $228 incl ship. Still don't know if I got my second. They still show 93 but something tells me a check at 6am means they haven't started weekend pulls. Good luck people.

 

[harpoon84]

I never mentioned memory, so I don't intend to go any googling... though for what its worth 1:1 is NOT the best, I've tested this myself, running the memory faster than the FSB yields a small performance gain.

The $133 figure is NOT the MSRP price, it's the official 1000qty price from Intel. Obviously stores will have to markup the price somewhat, for example the E4600 is about $140 - $150 depending on where you shop - this is what we can expect from the E7xxx chips if there is no price gouging involved.

I'm not 'upset' at all, but rather frustrated that for someone who has so much experience building PCs, you still don't understand the difference between the 1000 qty price (hence the $133 figure), and the retail price stores charge. The onus is on them to charge a fair price, but if people are stupid enough to buy the chips at ridiculously inflated prices then that is the perfect model of supply and demand I guess.

The shortage of the E8400 has nothing to do with the E7xxx chips. If there are still 45nm shortages by May/June, then that could be problem, but it's some way off yet and Intel is ramping up 45nm production.

 

[nonameo]

It fits in socket 775.

 

[Denithor]

Amazing that a 3MB chip will beat out a 4MB chip by 3% (after adjusting for clockspeed difference).

I think you've convinced me where my next $150 is going to go...just as soon as Intel gets off their butts and launches the e7300. Should be a nice upgrade from my e6400 with 2MB cache!

 

[harpoon84]

Originally posted by: Denithor
Amazing that a 3MB chip will beat out a 4MB chip by 3% (after adjusting for clockspeed difference).

I think you've convinced me where my next $150 is going to go...just as soon as Intel gets off their butts and launches the e7300. Should be a nice upgrade from my e6400 with 2MB cache!

Well at the same cache sizes 45nm is about ~5% faster than 65nm, so adjusting for the slight performance hit from the smaller cache ~3% sounds about right, not that amazing really if you think logically...

I'm also considering this chip as an upgrade for my E4400. To be honest, I don't really need the extra performance that much, but hey more performance can't hurt right , plus I like the power savings from the 45nm process.

 

[Denithor]

Originally posted by: harpoon84
To be honest, I don't really need the extra performance that much, but hey more performance can't hurt right , plus I like the power savings from the 45nm process.

I'm in exactly the same boat here, my e6400 is perfectly happy running 3GHz @ stock volts so I don't exactly need a performance boost...but I want one for both of the reasons you mentioned! Plus I will be able to sell off the e6400 to someone for probably $100 or so. Makes for a cheap upgrade any way you look at it.

 

[TheJian]

Originally posted by: harpoon84
I never mentioned memory, so I don't intend to go any googling... though for what its worth 1:1 is NOT the best, I've tested this myself, running the memory faster than the FSB yields a small performance gain.

The $133 figure is NOT the MSRP price, it's the official 1000qty price from Intel. Obviously stores will have to markup the price somewhat, for example the E4600 is about $140 - $150 depending on where you shop - this is what we can expect from the E7xxx chips if there is no price gouging involved.

I'm not 'upset' at all, but rather frustrated that for someone who has so much experience building PCs, you still don't understand the difference between the 1000 qty price (hence the $133 figure), and the retail price stores charge. The onus is on them to charge a fair price, but if people are stupid enough to buy the chips at ridiculously inflated prices then that is the perfect model of supply and demand I guess.

The shortage of the E8400 has nothing to do with the E7xxx chips. If there are still 45nm shortages by May/June, then that could be problem, but it's some way off yet and Intel is ramping up 45nm production.

It's obvious I just made a mistake with msrp and it should have been 1000qyt...I have accounts at Ingram (well, I hate them so don't use them...damn tiers pricing), ASI, Supercom etc. Cut me some slack I was getting lazy, look at the posts I'd been up all night (looking at your times so were you...rofl). I was bone tired and just trying to stay alive until Excalibur showed my other e3110 had shipped so I could cancel and order somewhere else if needed (which I didn't make it, but the cpu did so they're apparently still in people they've changed and show 97 still!...).

You hit on my whole point, we will be gouged. That was all I was saying. The e8400 is directly related in my mind. It shows us a taste of the future when the next bang for buck champ hits which is arguably any E7xxx chip if E8xxx chips haven't hit below $180-190. It would help if all the old chips are gone by the time this comes out but if not Intel will delay us and restrict chips (just like e8400, they aren't out, just restricted for old stock sell-offs). If this thing ends up going from $150 (msrp in my mind if Intel is charing $133/1000) to near $200 I'll be looking elsewhere for my nephews birthday in may :| Unfortunately there won't be a secret xeon model (if you could call the 3110 that) under another name floating around that we can pick up at non gouged pricing. That's depressing. No, I'll stick with maddening...I guess I'm a little upset about the prospect of getting burned a bit again (or more than a bit). I just paid an extra $20 for my dad's over my price 1 day earlier...ouch.

I suppose we'll all be running around like chickens with out head cut off trying to get one of these the first day they hit (before the automated price hikers on newegg etc force the price up). And yes, boatloads of us will be stupid enough to pay it. :roll: We've all proven it many times before...LOL.

Memory: I googled it now and ended up back here...LOL : http://forums.anandtech.com/me...AR_FORUMVIEWTMP=Linear
Check out Lopri's post. I haven't benched in a while but I can verify it was usually good 1:1 when I compared, but as he says there todays are better so maybe it's not quite the same these days. He does point out it's all over the map and gets worse at 4dimms. I also agree that the memory controller on Amd makes it not apply (nehalem should fix this divider crap too). Perhaps I overstated it regarding todays chipsets. I couldn't find anyone that benched them. I guess it says something that nobody cares.

 

[TheJian]

Originally posted by: Denithor

Originally posted by: harpoon84
To be honest, I don't really need the extra performance that much, but hey more performance can't hurt right , plus I like the power savings from the 45nm process.

I'm in exactly the same boat here, my e6400 is perfectly happy running 3GHz @ stock volts so I don't exactly need a performance boost...but I want one for both of the reasons you mentioned! Plus I will be able to sell off the e6400 to someone for probably $100 or so. Makes for a cheap upgrade any way you look at it.

I'm thinking I just bought my last chip for this board. I'm guessing you guys will too. I mean if you can hit 4ghz (more on water I hope with 3110) when it gets slow what will Intel come out with to beat it? I'm thinking I'll run at 3.6, ok maybe 3ghz for a while, there is nothing that slows my 4300@3ghz down. I just want to know it's there. You don't have to upgrade, just OC to whatever later (3.6 for the first bit, then 4.0+ later when forced and it's worthless so no fear). Intel will have to release a chip that can hit 5ghz+ to really make it worth not looking at some cheap nehalem at that point I think. I'm not even sure about that if it hits 4.2+ what's the point in upgrading to anything lower than 4.6 or so? Another spin or two and these penryns should hit 4.5 reliably across all models. Overclockers should then be getting 5ghz probably. Things to ponder I guess I already told my dad if either of us can't hit 3.6 close to default volts I'll ebay them 10 minutes later and try again...LOL.

 

[coldpower27]

Originally posted by: harpoon84

Originally posted by: Denithor
Amazing that a 3MB chip will beat out a 4MB chip by 3% (after adjusting for clockspeed difference).

I think you've convinced me where my next $150 is going to go...just as soon as Intel gets off their butts and launches the e7300. Should be a nice upgrade from my e6400 with 2MB cache!

Well at the same cache sizes 45nm is about ~5% faster than 65nm, so adjusting for the slight performance hit from the smaller cache ~3% sounds about right, not that amazing really if you think logically...

I'm also considering this chip as an upgrade for my E4400. To be honest, I don't really need the extra performance that much, but hey more performance can't hurt right , plus I like the power savings from the 45nm process.

This is quite tempting for me as well, but I am also considering waiting till the 45nm Core 2 Quad price drops in Q3 and jumping on the Q9550 then. It's just tough right now to decide.

 

[Denithor]

Originally posted by: coldpower27
This is quite tempting for me as well, but I am also considering waiting till the 45nm Core 2 Quad price drops in Q3 and jumping on the Q9550 then. It's just tough right now to decide.

My next "big" upgrade will be Nehalem sometime in 2009. This will just be a sideways move to reduce power consumption/temperature (and maybe clock a little higher while I'm at it) before switching platforms. My first quad will certainly be Nehalem not Penryn.

 

[SunSamurai]

Originally posted by: TheJian

Originally posted by: Denithor

Originally posted by: harpoon84
To be honest, I don't really need the extra performance that much, but hey more performance can't hurt right , plus I like the power savings from the 45nm process.

I'm in exactly the same boat here, my e6400 is perfectly happy running 3GHz @ stock volts so I don't exactly need a performance boost...but I want one for both of the reasons you mentioned! Plus I will be able to sell off the e6400 to someone for probably $100 or so. Makes for a cheap upgrade any way you look at it.

I'm thinking I just bought my last chip for this board. I'm guessing you guys will too. I mean if you can hit 4ghz (more on water I hope with 3110) when it gets slow what will Intel come out with to beat it? I'm thinking I'll run at 3.6, ok maybe 3ghz for a while, there is nothing that slows my 4300@3ghz down. I just want to know it's there. You don't have to upgrade, just OC to whatever later (3.6 for the first bit, then 4.0+ later when forced and it's worthless so no fear). Intel will have to release a chip that can hit 5ghz+ to really make it worth not looking at some cheap nehalem at that point I think. I'm not even sure about that if it hits 4.2+ what's the point in upgrading to anything lower than 4.6 or so? Another spin or two and these penryns should hit 4.5 reliably across all models. Overclockers should then be getting 5ghz probably. Things to ponder I guess I already told my dad if either of us can't hit 3.6 close to default volts I'll ebay them 10 minutes later and try again...LOL.

Sure, an old post, but what kind of rubbish is this?

I thought everyone learned by now that Mhz=!Mhz with new chip designs. You cannot measure the performance between different generations by Mhz. Else we would be at 50+ Ghz by now.