780G, err 790GX Northbridge
The biggest change AMD made in designing the 790GX is an increase in core clock speed from 500MHz to 700MHz. In essence, AMD is speed binning the 780G chipsets and those that pass end up with a 790GX silkscreen stamp. Due to the improved performance, the graphics processor also has a name change bump from Radeon HD 3200 to, wait for it, Radeon HD 3300.
The 790GX still utilizes a 55nm TSMC half-node process with transistor counts hitting 205 million. The Northbridge contains 26 lanes of PCI Express 2.0 connectivity with 16 lanes reserved for graphics via a single x16 link or a dual-x 8 CrossFire configuration. Six lanes are available for x1 links for expansion slots or onboard peripherals. AMD reserves the remaining four links for use in the A-Link Express II interface to the SB750 Southbridge.
Hybrid CrossFire returns but support is still limited to the HD 2400 and 3400 based cards. The improved clock speeds of the HD 3300 and performance cache allows a user with a HD 3470 to get performance near that of a HD 3650. The problem is that most HD 3470 cards are selling for around $50 and recent price reductions have brought the price of the HD3650 down to $65, in essence negating the previous price to performance advantages. We doubt anyone will take advantage of this feature since the chipset is targeted primarily for the performance sector where most users will take advantage of a single performance oriented card like the HD 4850 or a CrossFire configuration instead of Hybrid CF if the integrated graphics were not fast enough.
Sideport Memory returns energetically in the 790GX products, this time at the request of AMD. While an option on the 780G boards, AMD has persuaded the motherboard suppliers to include sideport memory in at least one 790GX product from their lineups. Along with this inclusion comes a new name, Performance Cache. The performance cache resides on a 16-bit memory interface that can utilize DDR2 or DDR3 memory. The majority of motherboard makers will use DDR3-1333 in a 128MB configuration. While performance between DDR2 and DDR3 is about the same in this particular application, DDR3 offers additional power savings.
Depending on the processor and HT Link utilized, sideport memory will generally improve performance across the board up to 5% - not just in games, but in video decoding capabilities as well. Besides performance improvements resulting from the local frame buffer instead of shared system memory, the new driver design allows improved Deep Sleep capabilities with the active frame buffers being stored in the performance cache.
As with the 780G, the 790GX AMD features Unified Video Decoder 2.0 (UVD 2.0) capabilities that include hardware acceleration for decoding VC-1, H.264 (AVC), WMV, and MPEG-2 sources up to 1080p resolutions. Advanced de-interlacing is available when using a Phenom processor. HDMI 1.3 compliance continues with xvYCC color space support and 4:2:2 capabilities. On the audio side, the HDMI interface offers support for 2-channel LPCM, Dolby Digital 5.1, or DTS 5.1. AMD still does not support Multi-Channel LPCM and it continues to be a thorn in our side.
Rounding out the video capabilities of the 790GX is analog output, DVI/HDMI interfaces, internal or external TMDS, and integrated DisplayPort capabilities. The HD 3300 features dual independent displays that allow resolution, refresh rates, and display data to be completely independent on the two display paths. AMD provides HDCP support with on-chip key storage for DisplayPort, DVI, or HDMI interfaces but it is regulated to a single interface during playback operations. ATI SurroundView is fully supported when utilizing a Radeon based discrete card that allows the system to drive up to four monitors.
The Radeon HD 3300 graphics processor features a DX10 compliant unified shader model 4.0 graphics core. This architecture contains 40 stream processors from two shader SIMDs. The core features 128-bit floating point precision for all operations, a command processor for reduced overhead, DXTC and 3Dc+ texture compression, texture support up to 8192x8192, and a fully associative texture and Z/stencil cache.
The single texture and ROP units are capable of handling four texels and pixels per-clock, respectively. This allows up to 16 texture fetches per clock cycle and up to 128 textures per pixel. The one modification AMD made compared to the RV610 core on the HD 2400 series is the vertex and texture caches are fully associative instead of separate. Other technical goodies include early Z-test, Re-Z, Z range optimization, and Fast Z-clear operation.
There are eight render targets (MRTs) with anti-aliasing support, along with lossless Z and stencil compression. The HD 3300 features multi-sample anti-aliasing (up to 4 samples per pixel), custom filter anti-aliasing, adaptive super-sampling and multi-sampling, along with Super AA when using Hybrid CrossFire. HDR rendering is supported for all anti-aliasing features, although like current HD 2400 products, the performance hit for using AA is significant in certain games.
Texture filtering features include 2x/4x/8x/16x adaptive anisotropic filtering modes with up to 128 taps per pixel support. Bicubic, sRGB, and percentage closer filtering are featured along with 128-bit floating point HDR support. Finally, the HD 3300 includes a programmable tessellation unit and an accelerated geometry shader path for geometry amplification.
Let’s move on the biggest improvement since sliced bread, well almost.
Facebook
Twitter
RSS
39 Comments
View All Comments
Janooo - Thursday, August 7, 2008 - link
Gary,you could be right when you say:
"Curiously enough, AMD cited "competitive concerns" as a reason why it would not disclose exactly what's going on with this new overclocking feature. We still cannot help but wonder if it is because AMD is going a little too far in the sacrifices, it is willing to make in the quest for higher clock speeds."
Though, another reason could be that AMD found something nice that would give Intel even higher edge if they knew what AMD did.
ZootyGray - Thursday, August 7, 2008 - link
"reason could be that AMD found something nice"Agree. And this same statement was uttered when ACC appeared in the SB750 review.
I would like to know - WHAT is the sacrifice? The apparently unqualified statement seems to cast doubt and nasty slur, and really should be explained, methinks.
ET - Thursday, August 7, 2008 - link
First of all, I'd like to say well done on letting us know of the issues and not rushing out a review like other sites. (Though I must say that having access to both fast and in depth reviews is the best of both worlds, so I'm also glad these other sites exist.)One thing which wasn't clear to me was: "vertex and texture caches are fully associative instead of separate." "Fully associative" has a meaning when it comes to caches, yet I imagine that's not what it's about. Sounds from the sentence like the caches have been unified (as opposed to separate).
Since according to the text this is the only departure from the 2400's architecture, it'd be nice to clarify this point.
quanta - Wednesday, August 6, 2008 - link
It seems the 'new' Advanced Clock Calibration is based on something that already existed in SB700 but was undocumented.[1] Based on the speculation, ACC could be the clock skew feature used by some other motherboards.That aside, RAID5 is a definite welcome addition, but unless it includes XOR offload, it really isn't much of an improvement. BTW, it is inexcusable for not having Ethernet support these days.
The claim of adding 2 specific 1.1 ports for compatibility reasons is confusing. USB 2.0 ports are 1.1 already compatible, just using slower speeds.
HyperFlash is a poorly implemented idea, especially when production Flash SSDs already exceed the maximum bandwidth of HyperFlash modules anyway. It is better off to just wait for Flash SSDs to mature.
[1] http://www.overclockers.com/tips01369/">http://www.overclockers.com/tips01369/
Goty - Wednesday, August 6, 2008 - link
It's possible, but I find it odd that a notoriously anti-AMD site (at least recently) like overclockers.com did absolutely no testing to confirm this hypothesis.chucky2 - Wednesday, August 6, 2008 - link
...to please give us 690G results along with it, for the benchmarks and power consumption. This way, we can see if 780G and/or 790GX, or one of the other competing solutions, is a worthwhile upgrade for us. One of the Gigabyte 690G boards would be perfect...Chuck
ZootyGray - Thursday, August 7, 2008 - link
And which manufacturer's 780G mobo would run with a Phenom 9950BE ? Is that possible?I am wondering how far the 780G comparison can go; since several mobos were destroyed back in April or May anand testing. (search The Three Amigos or "pop goes the mosfet"). Did the various manufacturers make any changes to replace the cheap mosfets they were using? 780G is a great N/B chip but some older(?) boards need lower-power cpu's. Other newer boards? - not sure yet. Still awaiting the update on all that. Did I miss something?
chucky2 - Thursday, August 7, 2008 - link
So when the review comes out, use the maximum CPU that the 690G board supports, and carry that forward to the 780G and 790GX boards.Also it'd be nice to see the CPU comparison on like minimum CPU needed to play 1080p content.
Also, best price/performance CPU out, cheapest dual core out, and older AM2 CPU's such as the 3600+ X2.
Show us what we can expect when it's not a Phenom running in there, i.e. give us information users - both existing and new - would find helpful in the review.
Chuck
yyrkoon - Wednesday, August 6, 2008 - link
Lets just say that Stability is the main reason why I purchase what I purchase name brand wise. Motherboards, memory, and power supplies are my main concerns and probably will be for a long time to come.Anyways, performance numbers etc are always good to know also, but I think it would be good if you guys at anandtech could put some 'stability figures' in your motherboard results as well. "This board performed well ..." would not work for those of us who want to know actual figures either. So maybe you guys can include a page, half a page or something dedicated strictly to stability ?
One of the main reasons why I am asking for this is as an example, I have an AM2 opteron system, as well as Core 2 Duo system, and while I love my AM2 system, and it *is* 'very stable', it still can not touch the rock solid stability of the Intel system. I do not pretend to know why, but I would love to figure out why . . . Both system have a board made by the same manufacturer, have used the same memory, video card(same company made both the GPU, and motherboard chipsets - nVidia)power supply, and hard drives. Hell, the Intel system has even been overclocked for 6-7 months now : /
Going from the above I am left at guessing that it is either 1) a hardware implementation, or 2) drivers. User error *is* a possibility, but I would like to think that since I have been fixing / integrating systems since the early 90's . . . well that I at least know half of what I am doing. That said, this 'issue' is still over my head.
Calin - Thursday, August 7, 2008 - link
It could be hardware implementation, it could be something that could be solved by a microcode/BIOS update, or maybe not. Also, there might be a difference between individual components (as similar as they are in theory), difference in general quality between the two mainboards, difference in component quality on boards (either by design or by chance/bad capacitors). It might be in the drivers, or just in things written in the processor errata which are not taken into account by the software.