I've never had a desktop. Still burning through my laptop as a poor college student. Have been thinking of getting the 1155 socket setup and start from scratch and build a pc. But this processor keeps sounding better and better. I think I'm going to just buy a cheapo 775 socket computer off ksl or craigslist and mess around with it, then once I get serious I'll invest in the 1150 setup. Sandy bridge and Ivy bridge are nice but might as well learn on a $50 setup.
not a bad idea if you are just getting into things, just don't make the mistake of getting an old system and then adding upgrades to it that will not work with your future 'real' build.
I tend to find really old stuff isn't any cheaper (except sometimes on ebay, but I don't trust them). Get something newer but low end and you get some nice features and upgrade options. If you're really budget-constained, maybe just get some peripherals to make your laptop a temporary desktop, though then you miss out on the fun of building your own machine :)
Totally agree with dananski, don't waste your time with 775, you will be paying too much for junk, Only time I build with 775 is when I am salvaging one of the many fried dells out there for friends, a new mobo and they are back in business, but for the price of the new mobo, it almost makes more sense to buy a cheapo 1155 board($50 bucks) and the awesome Celeron G530, its painful how quick that chip is compared to all the 775 chips out there. and you can easily pick it up for around $40.
That's my two cents, but in your area you will only find overpriced junk, especially on KSL.
I keep seeing 775 socket computers that have a passmark score at least 1.5X better than my old laptop for like $50. All I'd have to do is maybe add some ram or a small video card. I understand spending too much on a dead setup is quite foolish, but $50? Not a bad deal.
What are you smoking? Even the Q6600 which has been around for almost a decade still kills the Celeron G530 just look up the passmark score the Q6600 gets about 3k without overclocking while the G530 2.4k, again, what are you smoking pal?<Oh and the funny thing is the Q6600 isn't even the fastest 775 chip look at the Q9650 especially when that's overclocked it hangs with I-7's so please don't talk bad about the longest lasting socket Intel ever made, now THAT'S STAYING POWER!
128MB isn't enough to store assets, but it's enough for buffers. Depending on how bottlenecked GT3 is by memory operations, this could be quite interesting. The only thing I'd be concerned about at this point is whether 128MB is enough; for forward rendering games I imagine it will be, but in deferred renderers that G-buffer can get pretty large.
It should be a massive benefit anyway, if it's enough or not. And mid-term such a "L4 cache" might be what's desperately needed to signficantly increase IGP performance, without driving the number of memory channels and cost up. Might even be enough to get ahead of AMDs Fusion chips, if they can't get around the bandwidth problem.
Not all models will be shipping with 128MB, either. That's the highest-end models only. Given the performance predictions, I'd say 128MB probably won't show up in their ULV models at all.
If this is just normal DRAM used for the cache can someone explain to me how this can boost the performance? Unless it's very fast DRAM thats much faster than system memory.
Cache is typically much faster (though much smaller) than DRAM. It does not help with anything, but the things it will help with will see vast improvements over relying on 'slow' system memory.
Rumors from a year ago were that Intel was going to use a large number of cellphone/tablet/etc type ram chips and implement a silicon interconnect (much denser traces than with a PCB) inside the package to create an ultrawide data bus to compensate for the low clockrates of each individual chip.
With Haswell topping out at only 128Mb vs the 512mb-1gb being speculated about at the time it seems likely they've either scrapped the ultrawide bus, or are using a custom ram chip design with a much wider than normal databus instead of a commodity part.
The idea is the same as the eDRAM in the Xbox 360: store the frequently used buffers in the eDRAM to free up bandwidth to main memory while simultaneously giving access to those buffers a massive increase in performance. The performance benefits can be massive depending on how frequently those buffers are accessed and if they can contain the full buffer.
With 128 MB, that's enough for two full 32 bit buffers at 1920 x 1080 resolution.
Being on-package allows the bus to be much wider. If rumors of the silicon interposer are true, there could potentially be a 512 bit (or wider) bus. Also, without wires travelling off the package, latency and power consumption are hugely cut down.
And will this cache be available to the CPU as well as the onboard GPU? Whether by default or by special instructions, accessing it from the CPU could be useful.
I doubt it would have a huge effect on performance since most consumer workloads are I/O bound anyway and even L3 today isn't a huge performance impact in most consumer workloads but it would be interesting for some very large computationally complex data algorithms.
Decoupled ring from the CPU so I'd highly doubt it. It seems like it's a GPU-only thing.
I'm also assuming that the ULV parts won't have much of it in order to remain at that 17W TDP. There's just no way a 128mb cache can fit within that TDP.
I wouldn't put money on it being accessible as a cache by the CPU. Rather I **think** that this memory is directly accessible by the CPU as memory. Its actual cache functionality is managed by software at the driver level to determine what gets stored there.
I think it's because Haswell's ULV chips have a 10W TDP whereas Ivy Bridge ULV is 17W. To reducd the TDP so much and still see a perfomance gain is pretty impressive!
Why do you think it's not? AFAIK GT3 graphics are mobile only; and since that's the GPU that would be hurt most by the system ram bottleneck, I'd expect full power mobile parts to be among the 128mb models.
Well becasue Intel only states double the performance of IVY. Which a GT3 should already been able to achieve, With an additional 128MB cache i expect more then that.
It had better actually be some means of improving CPU performance...but given the modest gains from modest cache size increases, I'm guessing not...
I REALLY hope AMD can get back in the game here, because I'm REALLY tired of Intel blowing hundreds of millions of transistors on their worthless video no one should want. That alone has me almost wanting to just do AMD...
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
35 Comments
Back to Article
CaptainDoug - Wednesday, September 12, 2012 - link
I've never had a desktop. Still burning through my laptop as a poor college student. Have been thinking of getting the 1155 socket setup and start from scratch and build a pc. But this processor keeps sounding better and better. I think I'm going to just buy a cheapo 775 socket computer off ksl or craigslist and mess around with it, then once I get serious I'll invest in the 1150 setup. Sandy bridge and Ivy bridge are nice but might as well learn on a $50 setup.CaedenV - Wednesday, September 12, 2012 - link
not a bad idea if you are just getting into things, just don't make the mistake of getting an old system and then adding upgrades to it that will not work with your future 'real' build.CaptainDoug - Thursday, September 13, 2012 - link
Yep. I have done pretty much everything but actually get my hands into it. Very bored with that too...dananski - Wednesday, September 12, 2012 - link
I tend to find really old stuff isn't any cheaper (except sometimes on ebay, but I don't trust them). Get something newer but low end and you get some nice features and upgrade options. If you're really budget-constained, maybe just get some peripherals to make your laptop a temporary desktop, though then you miss out on the fun of building your own machine :)cyrusfox - Thursday, September 13, 2012 - link
Totally agree with dananski, don't waste your time with 775, you will be paying too much for junk, Only time I build with 775 is when I am salvaging one of the many fried dells out there for friends, a new mobo and they are back in business, but for the price of the new mobo, it almost makes more sense to buy a cheapo 1155 board($50 bucks) and the awesome Celeron G530, its painful how quick that chip is compared to all the 775 chips out there. and you can easily pick it up for around $40.That's my two cents, but in your area you will only find overpriced junk, especially on KSL.
CaptainDoug - Thursday, September 13, 2012 - link
I keep seeing 775 socket computers that have a passmark score at least 1.5X better than my old laptop for like $50. All I'd have to do is maybe add some ram or a small video card. I understand spending too much on a dead setup is quite foolish, but $50? Not a bad deal.holymoly cajole - Monday, April 8, 2013 - link
What are you smoking? Even the Q6600 which has been around for almost a decade still kills the Celeron G530 just look up the passmark score the Q6600 gets about 3k without overclocking while the G530 2.4k, again, what are you smoking pal?<Oh and the funny thing is the Q6600 isn't even the fastest 775 chip look at the Q9650 especially when that's overclocked it hangs with I-7's so please don't talk bad about the longest lasting socket Intel ever made, now THAT'S STAYING POWER!Ryan Smith - Wednesday, September 12, 2012 - link
128MB isn't enough to store assets, but it's enough for buffers. Depending on how bottlenecked GT3 is by memory operations, this could be quite interesting. The only thing I'd be concerned about at this point is whether 128MB is enough; for forward rendering games I imagine it will be, but in deferred renderers that G-buffer can get pretty large.MrSpadge - Wednesday, September 12, 2012 - link
It should be a massive benefit anyway, if it's enough or not. And mid-term such a "L4 cache" might be what's desperately needed to signficantly increase IGP performance, without driving the number of memory channels and cost up. Might even be enough to get ahead of AMDs Fusion chips, if they can't get around the bandwidth problem.Alexvrb - Wednesday, September 12, 2012 - link
Not all models will be shipping with 128MB, either. That's the highest-end models only. Given the performance predictions, I'd say 128MB probably won't show up in their ULV models at all.B3an - Wednesday, September 12, 2012 - link
If this is just normal DRAM used for the cache can someone explain to me how this can boost the performance? Unless it's very fast DRAM thats much faster than system memory.ImSpartacus - Wednesday, September 12, 2012 - link
Nothing is set in stone, but Mr. Shimpi said to expect buckets of bandwidth, so we probably shouldn't worry about effective speeds.CaedenV - Wednesday, September 12, 2012 - link
Cache is typically much faster (though much smaller) than DRAM. It does not help with anything, but the things it will help with will see vast improvements over relying on 'slow' system memory.CaedenV - Wednesday, September 12, 2012 - link
*everything, not anything... oopsDanNeely - Wednesday, September 12, 2012 - link
Rumors from a year ago were that Intel was going to use a large number of cellphone/tablet/etc type ram chips and implement a silicon interconnect (much denser traces than with a PCB) inside the package to create an ultrawide data bus to compensate for the low clockrates of each individual chip.With Haswell topping out at only 128Mb vs the 512mb-1gb being speculated about at the time it seems likely they've either scrapped the ultrawide bus, or are using a custom ram chip design with a much wider than normal databus instead of a commodity part.
extide - Wednesday, September 12, 2012 - link
That wasnt a rumor it was pure speculation.Kevin G - Wednesday, September 12, 2012 - link
The idea is the same as the eDRAM in the Xbox 360: store the frequently used buffers in the eDRAM to free up bandwidth to main memory while simultaneously giving access to those buffers a massive increase in performance. The performance benefits can be massive depending on how frequently those buffers are accessed and if they can contain the full buffer.With 128 MB, that's enough for two full 32 bit buffers at 1920 x 1080 resolution.
Kevin G - Wednesday, September 12, 2012 - link
Whoops, math fail. That'd be sixteen 32 bit buffers or eight 64 bit HDR buffers at 1920 x 1080 resolution.madmilk - Wednesday, September 12, 2012 - link
Being on-package allows the bus to be much wider. If rumors of the silicon interposer are true, there could potentially be a 512 bit (or wider) bus. Also, without wires travelling off the package, latency and power consumption are hugely cut down.jwilliams4200 - Wednesday, September 12, 2012 - link
Did you mean "Ivy Bridge GT2" instead of "Ivy Bridge GT4"? I don't think IvB has GT4 or GT3.tipoo - Wednesday, September 12, 2012 - link
Only the ULV parts do? Or is it optional for every type? Is the 2x with or without the cache?Ken g6 - Wednesday, September 12, 2012 - link
And will this cache be available to the CPU as well as the onboard GPU? Whether by default or by special instructions, accessing it from the CPU could be useful.MrSpadge - Wednesday, September 12, 2012 - link
Yeah, this sounds like something the CPU might love to use. And not only on ULV parts.aicom64 - Wednesday, September 12, 2012 - link
I doubt it would have a huge effect on performance since most consumer workloads are I/O bound anyway and even L3 today isn't a huge performance impact in most consumer workloads but it would be interesting for some very large computationally complex data algorithms.tipoo - Wednesday, September 12, 2012 - link
Yeah, it's would be a bit like IBMs Power7 in that regard, with a huge on die memory chip.mrdude - Wednesday, September 12, 2012 - link
Decoupled ring from the CPU so I'd highly doubt it. It seems like it's a GPU-only thing.I'm also assuming that the ULV parts won't have much of it in order to remain at that 17W TDP. There's just no way a 128mb cache can fit within that TDP.
Kevin G - Wednesday, September 12, 2012 - link
I wouldn't put money on it being accessible as a cache by the CPU. Rather I **think** that this memory is directly accessible by the CPU as memory. Its actual cache functionality is managed by software at the driver level to determine what gets stored there.winterlude - Wednesday, September 12, 2012 - link
This doesn't make sense to me. Do you mean 130% faster? Isn't Haswell's GT3 already double Ivy's best offerings?andykins - Thursday, September 13, 2012 - link
I think it's because Haswell's ULV chips have a 10W TDP whereas Ivy Bridge ULV is 17W. To reducd the TDP so much and still see a perfomance gain is pretty impressive!LuckyKnight - Thursday, September 13, 2012 - link
Question is.. Have they fixed 23.967 hz outputiwod - Thursday, September 13, 2012 - link
Why isn't this on Mobile as well? It could truly eliminate the need for discrete graphics.DanNeely - Thursday, September 13, 2012 - link
Why do you think it's not? AFAIK GT3 graphics are mobile only; and since that's the GPU that would be hurt most by the system ram bottleneck, I'd expect full power mobile parts to be among the 128mb models.iwod - Sunday, September 16, 2012 - link
Well becasue Intel only states double the performance of IVY. Which a GT3 should already been able to achieve, With an additional 128MB cache i expect more then that.AnTech - Thursday, September 13, 2012 - link
When quad-core ULV?Wolfpup - Monday, October 1, 2012 - link
It had better actually be some means of improving CPU performance...but given the modest gains from modest cache size increases, I'm guessing not...I REALLY hope AMD can get back in the game here, because I'm REALLY tired of Intel blowing hundreds of millions of transistors on their worthless video no one should want. That alone has me almost wanting to just do AMD...