Pity they didn't have a warehouse full of 1GB sticks to get rid of as well, then they could do a Spinal Tap edition that went up to 11
Pity they didn't have a warehouse full of 1GB sticks to get rid of as well, then they could do a Spinal Tap edition that went up to 11
I'd have to check the BIOS settings - it does have sideport but I can't remember how it's using it
When I first got the board I was running the graphics *just* on the sideport (to preserve as much of the 2GB RAM as possible for the system), so 128MB, 32bit, DDR3-1333 memory interface, and for older games and windows desktop it worked fine I think it's probably currently on sideport + shared memory, so potentially it's got a 96bit interface to run from.
Then why not shove 2 in each box and have 12GB in dual channel...?
Now *that* would've made sense
Engadget reviewed a trinity notebook from HP (Envy Sleekbook 6z)
http://www.engadget.com/2012/07/13/h...ook-6z-review/
Our 6z comes with a dual-core AMD A6-4455M processor clocked at 2.1GHz with 4GB of RAM, Radeon HD 7500G graphics and a 500GB hard drive.It's pretty clear that HP's new Sleekbook line is targeted toward those who fancy a responsive, high-performing thin-and-light but don't want to shell out a grand or more for an Ultrabook. We appreciate the company's effort to temper features with price, and while the 6z is capable enough to handle your everyday computing needs, the underpowered AMD APU isn't a great performer -- and you can do better without spending much more, as even the Envy Ultrabook 6t starts at just $800. For what it's worth, though, folks who do opt for the Intel-approved version will enjoy the same understated design and comfortable keyboard (and hopefully soon-to-be-refined touchpad), so it's definitely worth looking at HP's Envy line anyway if the performance of this particular unit is the main thing giving you pause.
All 125w but one !!! Also a strange turbo mode for the FX8300 that goes down in speed ???? Mistake ?
http://lenzfire.com/2012/07/am3-sock...-specs-leaked/
KeyboardDemon (15-07-2012)
Thanks for the article, I'm expecting there to be a number of leaked results like this appearing over the next few months leading up to the actual launch, but as it said in the last paragraph:
Strangely they tested these CPU with a native FSB speed of 1866 using ram set at 1333, as has been previously demonstrated by Cat in an earlier post. I wish I could understand the logic behind that, but at least they give the information quite clearly on the Benchmark tables.In WinRARx64 and PCMark 7 benchmarks, Piledriver modules performed almost 60% better than Llano APUs. By analysing the CINEBench R11.5 x64 single thread benchmark, we can say that AMD had worked hard and improved the IPC of Piledriver modules to about 7%, which is half of the promised 15%. But still, we cannot come to a conclusion by comparing a single benchmark, moreover, it’s uncertain how the 4 Piledriver modules in AMD Vishera CPU would perform, as Vishera has 8MB additional cache memory compared to Trinity APUs.
It looks like Trinity for the desktop is being released in October together with Vishera:
http://www.techpowerup.com/169100/AM...ot-Launch.html
It seems a really weird move especially with the CPUs already being released in prebuilt desktops.
The IB Core i3 will be released by then,with better CPU performance than the SB Core i3 CPUs and one model will have the HD4000 IGP. They should be releasing the CPUs sooner,so any comparisons are made to the SB Core i3 CPUs. No doubt the Trinity A10 and probably the A8 IGPs will beat the HD4000,but still they look more impressive if compared to the HD3000.
Last edited by CAT-THE-FIFTH; 16-07-2012 at 12:32 PM.
It's almost like AMD don't want to sell any of the new parts - the only logical explanation I can see is if OEMs are demanding far more inventory than AMD can provide, but then you'd expect shops to be full of Trinity-based pre-built desktops and laptops, and AFAIK that's not (currently) the case.
Other possible explanations anyone? I suppose there could be issues with the socket? Motherboard manufacturers not ready yet (although I thought some manufacturers had already shown retail-ready FM2 motherboards)? Chipset issues? But if that were the case then they couldn't be supplying to OEMs...
No, there's only two things that make any sense: either supply issues as above, or they have massive stocks of Llano that they're hoping will shift if they delay Trinity. And if it's the latter, I think that's both poor planning and poor reading of the market, tbh...
Fudzilla seems to get confused these days. Rumours of quad core specs, but are they really changing the amount of cache?
http://www.fudzilla.com/home/item/27...parts-in-q4-12
If these parts are 8MB L3 + 4x1MB L2 then perhaps they are inconsistent on whether they include L2 in the total, otherwise some of those parts are 4MB L3 and that would be a bit sucky.
Seems that Piledriver release means a last buy notification for bulldozer.
http://www.fudzilla.com/home/item/27...ies-in-q4-2012
Possibly they'll be able to produce the bottom bin parts cheaper if they're cutting L3 cache out as well? A 2-module piledriver FX for < £80 could be pretty tasty as the base to a value gaming PC. And if 4-module parts have to share 8MB L3 cache, then potentially 2-module parts won't take too much of a hit from sharing half that? It could be that they think IPC and cache hit rate has improved sufficiently that they can get away with less cache ... after all, 8MB total cache is the same as quad-core Phenom II...
Only benchmarks will tell, I guess.
ISTR the Phenom I was available in different L3 cache sizes, so they have played that game before.
If they did a proper 2 module/4 core release of the silicon then I would have thought that could be a lot cheaper to make, and yes I would expect them to scale down the L3 on such a chip.
I *thought* PI was only available with 2MB L3 cache, and that was one of the reasons it sucked When PII first launched there were some OEM versions with reduced L3 cache, but they vanished pretty quickly (presumably due to yields being good enough to make fusing off cache unnecessary).
I'd love to think that AMD were going to have a separate 2-module die for Vishera to make the entry level processors cheaper (and less power hungry?), but I'm sceptical - there was meant to be a separate dual-core die for Llano but I don't believe that ever materialised either.
Ah yes, Phenom and early Phenom II must have gotten mixed up in my brain. Only Phenom in our house was the one branded Athlon X2 7750 with 2 cores disabled, such was the confusion on branding with these things. Not a bad chip, did run a bit hot though.
Still, reading on the Register today about the Power 7 chip on 45nm having 32MB of cache makes all of this look a bit pants.
HP Sleekbook 6z review:
http://www.engadget.com/2012/07/13/h...ook-6z-review/
It costs around $600.
Sadly,a fail of a review. The laptop only has a single SODIMM meaning the APU is crippled. Moreover,they ignore the fact that you can get an A10-4655M with 6GB of DDR3 or 8GB of DDR3 for around $750 to $800:
http://www.shopping.hp.com/webapp/sh....do#anchor-top
Instead they say and go and get one the Envy Ultrabook 6t instead for around $800(since $200 is not much more). However,this comes with a SB ULV Core i5 or an IB ULV Core i5 for $25 more and single channel memory. When you bump up the memory to 6GB to 8GB of dual channel memory you are looking at another $50 to $100 more.
There are currently 28 users browsing this thread. (0 members and 28 guests)