VGleaks: World Exclusive Orbis Unveiled!

This topic is locked from further discussion.

#51 Posted by clyde46 (46997 posts) -

are these the actual specs? There's alot of mixed reactions going around.

ReadingRainbow4
They are looking more and more likely to be the real deal now.
#52 Posted by Stevo_the_gamer (43157 posts) -

- Internal mass storage

  • One SKU at launch: 500 GB HDD
  • There may also be a Flash drive SKU in the future

Blazerdt47

Don't care about the rest of the rumor poo but I really want this to be true. I think Nintendo made the best decision with having one SKU last generation for the longest time, and I really hope Microsoft and Sony can follow suit with that same line of thinking. I hate this differentiation between a "basic" and "premium," or like last generation... arcade and pro.

#53 Posted by ReadingRainbow4 (14737 posts) -

[QUOTE="ReadingRainbow4"]

are these the actual specs? There's alot of mixed reactions going around.

clyde46

They are looking more and more likely to be the real deal now.

Yeah, reading through the neogaf thead and the Durango specs seem to be along the same lines although weaker.

Kinect Box is looking more and more likely.

#54 Posted by faizan_faizan (7867 posts) -
[QUOTE="ReadingRainbow4"]

are these the actual specs? There's alot of mixed reactions going around.

clyde46
They are looking more and more likely to be the real deal now.

Long time no post.
#55 Posted by clyde46 (46997 posts) -
[QUOTE="clyde46"][QUOTE="ReadingRainbow4"]

are these the actual specs? There's alot of mixed reactions going around.

faizan_faizan
They are looking more and more likely to be the real deal now.

Long time no post.

7 day suspension. "Forum invasion behavior"
#56 Posted by clyde46 (46997 posts) -

[QUOTE="clyde46"][QUOTE="ReadingRainbow4"]

are these the actual specs? There's alot of mixed reactions going around.

ReadingRainbow4

They are looking more and more likely to be the real deal now.

Yeah, reading through the neogaf thead and the Durango specs seem to be along the same lines although weaker.

Kinect Box is looking more and more likely.

Maybe, I personally think MS would lose tons of market share they just went after Nintendo's pot of gold. Casuals don't really buy all that many games.
#57 Posted by faizan_faizan (7867 posts) -
[QUOTE="clyde46"][QUOTE="faizan_faizan"][QUOTE="clyde46"] They are looking more and more likely to be the real deal now.

Long time no post.

7 day suspension. "Forum invasion behavior"

I see.
#58 Posted by tormentos (19277 posts) -
[QUOTE="RR360DD"]This thread sucks until tormento turns up *waits*Tessellation
he was celebrating the 192GB/s of bandwidth and now that they changed it to 176GB/s is around durango's territory,of course he is going to deny these rumors because they don't make his future plastic girlfriend any better :cool:

Actually i wasn't celebrating but if you most know the PS4 one still quite faster,the 720 on is 68GB/s + 102GB/s but that doesn't mean they should be accounted like 170GB/s,go search on Beyond or Neogaf you can't combine those numbers,is 68 from one point to another and 102 from one point to another.. Unlike the PS4 176GB/s one.
#59 Posted by tormentos (19277 posts) -
These numbers don't add up for me. 1.6 billion triangles per second seems low while xenon did 500 million per sec. Sony claimed the ps3 could do 2.0 teraflops and this is 1.8? Would it really be that expensive to overclock that CPU to atleast 2.0ghz? Audio processor? Doesn't the CPU usually handle that. Where is ronvalencia when u need him.GotNugz
The RSX did 275 almost half of the 360 ones and we know how that story end,theory numbers means sh**.
#60 Posted by Tessellation (8803 posts) -
[QUOTE="Tessellation"][QUOTE="RR360DD"]This thread sucks until tormento turns up *waits*tormentos
he was celebrating the 192GB/s of bandwidth and now that they changed it to 176GB/s is around durango's territory,of course he is going to deny these rumors because they don't make his future plastic girlfriend any better :cool:

Actually i wasn't celebrating but if you most know the PS4 one still quite faster,the 720 on is 68GB/s + 102GB/s but that doesn't mean they should be accounted like 170GB/s,go search on Beyond or Neogaf you can't combine those numbers,is 68 from one point to another and 102 from one point to another.. Unlike the PS4 176GB/s one.

we don't know anything about either console,these are just rumors for all we know these two speeds can work together,maybe not? or MS worked it around to fix the bandwidth,remember how they increased the xbox 360 RAM quite late,we don't know anything to base things of just speculations.
#61 Posted by TheEroica (14110 posts) -

I havent been around SW the past few days and honestly, I think its because of the rumors... Im excited like everyone else, but it just feels so useless to specualte considering it means nothing till the company unveils it. We're all stuck in the weird haze between generations.

#62 Posted by TheXFiles88 (1040 posts) -

[QUOTE="Tessellation"][QUOTE="RR360DD"]This thread sucks until tormento turns up *waits*tormentos
he was celebrating the 192GB/s of bandwidth and now that they changed it to 176GB/s is around durango's territory,of course he is going to deny these rumors because they don't make his future plastic girlfriend any better :cool:

Actually i wasn't celebrating but if you most know the PS4 one still quite faster,the 720 on is 68GB/s + 102GB/s but that doesn't mean they should be accounted like 170GB/s,go search on Beyond or Neogaf you can't combine those numbers,is 68 from one point to another and 102 from one point to another.. Unlike the PS4 176GB/s one.

You know, the DDRx DRAM has already been outclassed by a new revolutionary DRAM which called "Hybrid Memory Cube" It is a multi-chip module which works in parallel with the multi-core processors. The advantages are enormous over the DDRx. It costs less, 15x more performance, consumes less power and takes less space than a typical DDR3. If Microsoft is going to use this new HMC in its new Xbox (I don't think why they shouldn't), it will be a game changer since Microsoft is currently 1 of the 10 members of the HMC consortium. Just wait for the final specs after E3...;)

"Hybrid Memory Cube is a revolutionary innovation in DRAM memory architecture that sets a new standard for memory performance, power consumption and cost."

- HMC Combines high-speed logic process technology with a stack of through-silicon-via (TSV) bonded memory die.

- HMC delivers dramatic improvements in performance, breaking through the memory wall and enabling dramatic performance and bandwidth improvements -a single HMC can provide more than 15x the performance of a DDR3 module.

- The revolutionary architecture of HMC is exponentially more efficient than current memory, utilizing 70% less energy per bit than DDR3 DRAM technologies..

- Hybrid Memory Cube's increased density per bit and reduced form factor contribute to lower total cost of ownership, by allowing more memory into each machine and using nearly 90% less space than today's RDIMMs.

- With performance levels that break through the memory wall, Hybrid Memory Cube represents the key to extending network system performance to push through the challenges of new 100G and 400G infrastructure growth. Eventually, HMC will drive exascale CPU system performance growth for next generation HPC systems.

- Whereas DDR4 represents an evolutionary standard, HMC is a revolutionary technology that is a complete paradigm shift from current memory architectures.

- Hybrid Memory Cube will redefine memory. By advancing past the traditional DRAM system, HMC is setting a new standard of memory that can keep up with the advancements of CPUs and GPUs.

- Hybrid Memory Cube could be an absolute game changer for applications ranging from high performance computing to consumer technologies like tablets and graphics cards that value a combination of form factor, energy and bandwidth.

http://www.hybridmemorycube.org/technology.html

"Micron Readies Hybrid Memory Cube for Debut"

Hybrid Memory Cube (HMC) technology, a multi-chip module (MCM) that aims to address one of the biggest challenges in high performance computing: scaling the memory wall.

Memory architectures haven't kept pace with the bandwidth requirements of multicore processors. As microprocessor speeds out-accelerated DRAM memory speeds, a bottleneck developed that is referred to as the memory wall. Stacked memory applications, however, enable higher memory bandwidth.

The Hybrid Memory Cube (HMC) is a new memory architecture that combines a high-speed logic layer with a stack of through-silicon-via (TSV) bonded memory die that enables impressive advantages over current technology. According to company figures, a single HMC offers a 15x performance increase and uses 70 percent less energy per bit when compared to DDR3 memory, and takes up 90 percent less space than today's RDIMMs. The Cube is also scalable per application, which is not possible with DDR3 and DDR4. System designers have the option of employing the HMC as near memory for best performance or in a scalable module form factor, as far memory, for optimum power efficiency. This is a huge leap forward from a technology perspective, noted Graham, compared to DDRx and other boutique memory products that are out there.

http://www.hpcwire.com/hpcwire/2013-...for_debut.html

#63 Posted by SUD123456 (4520 posts) -

[QUOTE="tormentos"][QUOTE="Tessellation"] he was celebrating the 192GB/s of bandwidth and now that they changed it to 176GB/s is around durango's territory,of course he is going to deny these rumors because they don't make his future plastic girlfriend any better :cool:TheXFiles88

Actually i wasn't celebrating but if you most know the PS4 one still quite faster,the 720 on is 68GB/s + 102GB/s but that doesn't mean they should be accounted like 170GB/s,go search on Beyond or Neogaf you can't combine those numbers,is 68 from one point to another and 102 from one point to another.. Unlike the PS4 176GB/s one.

You know, the DDRx DRAM has already been outclassed by a new revolutionary DRAM which called "Hybrid Memory Cube" It is a multi-chip module which works in parallel with the multi-core processors. The advantages are enormous over the DDRx. It costs less, 15x more performance, consumes less power and takes less space than a typical DDR3. If Microsoft is going to use this new HMC in its new Xbox (I don't think why they shouldn't), it will be a game changer since Microsoft is currently 1 of the 10 members of the HMC consortium. Just wait for the final specs after E3...;)

"Hybrid Memory Cube is a revolutionary innovation in DRAM memory architecture that sets a new standard for memory performance, power consumption and cost."

- HMC Combines high-speed logic process technology with a stack of through-silicon-via (TSV) bonded memory die.

- HMC delivers dramatic improvements in performance, breaking through the memory wall and enabling dramatic performance and bandwidth improvements -a single HMC can provide more than 15x the performance of a DDR3 module.

- The revolutionary architecture of HMC is exponentially more efficient than current memory, utilizing 70% less energy per bit than DDR3 DRAM technologies..

- Hybrid Memory Cube's increased density per bit and reduced form factor contribute to lower total cost of ownership, by allowing more memory into each machine and using nearly 90% less space than today's RDIMMs.

- With performance levels that break through the memory wall, Hybrid Memory Cube represents the key to extending network system performance to push through the challenges of new 100G and 400G infrastructure growth. Eventually, HMC will drive exascale CPU system performance growth for next generation HPC systems.

- Whereas DDR4 represents an evolutionary standard, HMC is a revolutionary technology that is a complete paradigm shift from current memory architectures.

- Hybrid Memory Cube will redefine memory. By advancing past the traditional DRAM system, HMC is setting a new standard of memory that can keep up with the advancements of CPUs and GPUs.

- Hybrid Memory Cube could be an absolute game changer for applications ranging from high performance computing to consumer technologies like tablets and graphics cards that value a combination of form factor, energy and bandwidth.

http://www.hybridmemorycube.org/technology.html

"Micron Readies Hybrid Memory Cube for Debut"

Hybrid Memory Cube (HMC) technology, a multi-chip module (MCM) that aims to address one of the biggest challenges in high performance computing: scaling the memory wall.

Memory architectures haven't kept pace with the bandwidth requirements of multicore processors. As microprocessor speeds out-accelerated DRAM memory speeds, a bottleneck developed that is referred to as the memory wall. Stacked memory applications, however, enable higher memory bandwidth.

The Hybrid Memory Cube (HMC) is a new memory architecture that combines a high-speed logic layer with a stack of through-silicon-via (TSV) bonded memory die that enables impressive advantages over current technology. According to company figures, a single HMC offers a 15x performance increase and uses 70 percent less energy per bit when compared to DDR3 memory, and takes up 90 percent less space than today's RDIMMs. The Cube is also scalable per application, which is not possible with DDR3 and DDR4. System designers have the option of employing the HMC as near memory for best performance or in a scalable module form factor, as far memory, for optimum power efficiency. This is a huge leap forward from a technology perspective, noted Graham, compared to DDRx and other boutique memory products that are out there.

http://www.hpcwire.com/hpcwire/2013-...for_debut.html

Thanks for the links. Never heard of it till now, so it was interesting to read.

Also, to pour some gasoline on the fire I found this link as well:

It will have applications in servers but the most obvious application is in games. Faster memory together with faster processors will make for smoother high-resolution graphics. That is where you will notice the most change. It won't make any noticeable difference for simple things like accounting or word processing.

Linky

#64 Posted by lamprey263 (24667 posts) -
seems more legit since TC used a bigger font
#65 Posted by layton2012 (3653 posts) -
Another rumor that cows are going to brag about as if it is fact, I will wait until official specs before getting hyped.
#66 Posted by HaloPimp978 (7315 posts) -

Rumors, Rumors, Rumors :roll:

#67 Posted by clyde46 (46997 posts) -

Rumors, Rumors, Rumors :roll:

HaloPimp978
This is what happens when the the next generation of consoles are inbound.
#68 Posted by mitu123 (154467 posts) -

This system's specs is looking unimpressive with each rumor...

#69 Posted by ItsEvolution (2593 posts) -

Don't care. The company with the most games I want to play this fall is the one that will get my money for a new console. Simple as that.

#70 Posted by killzowned24 (7345 posts) -

Sounds like specs got better for Orbis :D

Orbis not only has a more powerful GPU,but now also has 4 compute units dedicated for whatever like cloth physics and such :P

#71 Posted by campzor (34932 posts) -
I cant understand tech mumbo jumbo.. just tell me with dbz power levels how it compares to durango (rumored) + wii u + ps3/360
#72 Posted by loco145 (10804 posts) -

$599.

#73 Posted by Heil68 (46127 posts) -

$599.

loco145
Day 1.
#74 Posted by ItsEvolution (2593 posts) -

$599.

loco145
If Sony expects to sell any systems, it can't be more than $399. It just can't be.
#75 Posted by Zophar87 (4351 posts) -

I'll probably be Wii U, PS4, and PC this gen.

#76 Posted by princeofshapeir (14644 posts) -
800 MHz GPU lel
#77 Posted by ronvalencia (15265 posts) -

UPDATE: some people is confused about the GPU, here you have more info about it:

Each CU contains dedicated:

- ALU (32 64-bit operations per cycle)

- Texture Unit

- L1 data cache

- Local data share (LDS)

About 14 + 4 balance:

- 4 additional CUs (410 Gflops) extra ALU as resource for compute

- Minor boost if used for rendering

Dual Shader Engines:

- 1.6 billion triangles/s, 1.6 billion vertices/s

18 Texture units

- 56 billion bilinear texture reads/s

- Can utilize full memory bandwith

8 Render backends:

- 32 color ops/cycle

- 128 depth ops/cycle

- Can utilize full memory bandwith

All this info issubject to changein the future by Sony.

Blazerdt47
On AMD Tahiti's CU, it combines two 32bit ALUs to form one 64bit ALU. AMD Tahiti has a ratio 1:4 on 64bit vs 32bit i.e. 16 ALUs with 64bit mode per CU. From AMD's dev forum, the next flagship GCN has a ratio 1:2 on 64bit vs 32bit i.e. 32 ALUs with 64bit mode per CU. It looks like PS4 would be based on next flagship GCN's CU but cut down i.e. reduced from 40 CU to 18 CU.
#78 Posted by GioVela2010 (4267 posts) -
Day 1, my GTX 670 will be raped
#79 Posted by ronvalencia (15265 posts) -

Day 1, my GTX 670 will be rapedGioVela2010

On 64bit shader, Liverpool GPU would murder it.

In terms of 64bit shader compute, Liverpool = 32 x 18 CU = 576 stream processors @ 800 Mhz

Tahiti XT2 = 16 x 32 CU = 512 stream processors @ 1Ghz

Tahiti XT = 16 x 32 CU = 512 stream processors @ 925 Mhz

It's x1900/Xenos (full speed 32bit compute) vs G7x again....

PS; Atm, I don't see the benefit of 64bit compute for games.

#80 Posted by Heirren (18089 posts) -
[QUOTE="loco145"]

$599.

ItsEvolution
If Sony expects to sell any systems, it can't be more than $399. It just can't be.

The market is different now. Personally I think people are more willing, now, to spend higher amounts than in the past. People spend spend $700 on iPads that are just really expensive toys that begin phasing out just after a few years.
#81 Posted by TheXFiles88 (1040 posts) -

[QUOTE="TheXFiles88"]

[QUOTE="tormentos"] Actually i wasn't celebrating but if you most know the PS4 one still quite faster,the 720 on is 68GB/s + 102GB/s but that doesn't mean they should be accounted like 170GB/s,go search on Beyond or Neogaf you can't combine those numbers,is 68 from one point to another and 102 from one point to another.. Unlike the PS4 176GB/s one.SUD123456

You know, the DDRx DRAM has already been outclassed by a new revolutionary DRAM which called "Hybrid Memory Cube" It is a multi-chip module which works in parallel with the multi-core processors. The advantages are enormous over the DDRx. It costs less, 15x more performance, consumes less power and takes less space than a typical DDR3. If Microsoft is going to use this new HMC in its new Xbox (I don't think why they shouldn't), it will be a game changer since Microsoft is currently 1 of the 10 members of the HMC consortium. Just wait for the final specs after E3...;)

"Hybrid Memory Cube is a revolutionary innovation in DRAM memory architecture that sets a new standard for memory performance, power consumption and cost."

- HMC Combines high-speed logic process technology with a stack of through-silicon-via (TSV) bonded memory die.

- HMC delivers dramatic improvements in performance, breaking through the memory wall and enabling dramatic performance and bandwidth improvements -a single HMC can provide more than 15x the performance of a DDR3 module.

- The revolutionary architecture of HMC is exponentially more efficient than current memory, utilizing 70% less energy per bit than DDR3 DRAM technologies..

- Hybrid Memory Cube's increased density per bit and reduced form factor contribute to lower total cost of ownership, by allowing more memory into each machine and using nearly 90% less space than today's RDIMMs.

- With performance levels that break through the memory wall, Hybrid Memory Cube represents the key to extending network system performance to push through the challenges of new 100G and 400G infrastructure growth. Eventually, HMC will drive exascale CPU system performance growth for next generation HPC systems.

- Whereas DDR4 represents an evolutionary standard, HMC is a revolutionary technology that is a complete paradigm shift from current memory architectures.

- Hybrid Memory Cube will redefine memory. By advancing past the traditional DRAM system, HMC is setting a new standard of memory that can keep up with the advancements of CPUs and GPUs.

- Hybrid Memory Cube could be an absolute game changer for applications ranging from high performance computing to consumer technologies like tablets and graphics cards that value a combination of form factor, energy and bandwidth.

http://www.hybridmemorycube.org/technology.html

"Micron Readies Hybrid Memory Cube for Debut"

Hybrid Memory Cube (HMC) technology, a multi-chip module (MCM) that aims to address one of the biggest challenges in high performance computing: scaling the memory wall.

Memory architectures haven't kept pace with the bandwidth requirements of multicore processors. As microprocessor speeds out-accelerated DRAM memory speeds, a bottleneck developed that is referred to as the memory wall. Stacked memory applications, however, enable higher memory bandwidth.

The Hybrid Memory Cube (HMC) is a new memory architecture that combines a high-speed logic layer with a stack of through-silicon-via (TSV) bonded memory die that enables impressive advantages over current technology. According to company figures, a single HMC offers a 15x performance increase and uses 70 percent less energy per bit when compared to DDR3 memory, and takes up 90 percent less space than today's RDIMMs. The Cube is also scalable per application, which is not possible with DDR3 and DDR4. System designers have the option of employing the HMC as near memory for best performance or in a scalable module form factor, as far memory, for optimum power efficiency. This is a huge leap forward from a technology perspective, noted Graham, compared to DDRx and other boutique memory products that are out there.

http://www.hpcwire.com/hpcwire/2013-...for_debut.html

Thanks for the links. Never heard of it till now, so it was interesting to read.

Also, to pour some gasoline on the fire I found this link as well:

It will have applications in servers but the most obvious application is in games. Faster memory together with faster processors will make for smoother high-resolution graphics. That is where you will notice the most change. It won't make any noticeable difference for simple things like accounting or word processing.

Linky

Yeah and this quote:"The Cube is also scalable per application....System designers have the option of employing the HMC as near memory for best performance or in a scalable module form factor, as far memory, for optimum power efficiency"is very interesting since it goes in lieu with that leaked "Scalable Console patent". Here is the chart of how the HMC vs. DRR3 & DDR4:

1

#82 Posted by ronvalencia (15265 posts) -
[QUOTE="princeofshapeir"]800 MHz GPU lel

It's 64bit compute would murder Geforce GTX 680. You are going to need Geforce Kepler K20 (acts like "8800 GTX" hammer but for 64bit compute).
#83 Posted by ronvalencia (15265 posts) -

[QUOTE="SUD123456"]

[QUOTE="TheXFiles88"]

You know, the DDRx DRAM has already been outclassed by a new revolutionary DRAM which called "Hybrid Memory Cube" It is a multi-chip module which works in parallel with the multi-core processors. The advantages are enormous over the DDRx. It costs less, 15x more performance, consumes less power and takes less space than a typical DDR3. If Microsoft is going to use this new HMC in its new Xbox (I don't think why they shouldn't), it will be a game changer since Microsoft is currently 1 of the 10 members of the HMC consortium. Just wait for the final specs after E3...;)

"Hybrid Memory Cube is a revolutionary innovation in DRAM memory architecture that sets a new standard for memory performance, power consumption and cost."

- HMC Combines high-speed logic process technology with a stack of through-silicon-via (TSV) bonded memory die.

- HMC delivers dramatic improvements in performance, breaking through the memory wall and enabling dramatic performance and bandwidth improvements -a single HMC can provide more than 15x the performance of a DDR3 module.

- The revolutionary architecture of HMC is exponentially more efficient than current memory, utilizing 70% less energy per bit than DDR3 DRAM technologies..

- Hybrid Memory Cube's increased density per bit and reduced form factor contribute to lower total cost of ownership, by allowing more memory into each machine and using nearly 90% less space than today's RDIMMs.

- With performance levels that break through the memory wall, Hybrid Memory Cube represents the key to extending network system performance to push through the challenges of new 100G and 400G infrastructure growth. Eventually, HMC will drive exascale CPU system performance growth for next generation HPC systems.

- Whereas DDR4 represents an evolutionary standard, HMC is a revolutionary technology that is a complete paradigm shift from current memory architectures.

- Hybrid Memory Cube will redefine memory. By advancing past the traditional DRAM system, HMC is setting a new standard of memory that can keep up with the advancements of CPUs and GPUs.

- Hybrid Memory Cube could be an absolute game changer for applications ranging from high performance computing to consumer technologies like tablets and graphics cards that value a combination of form factor, energy and bandwidth.

http://www.hybridmemorycube.org/technology.html

"Micron Readies Hybrid Memory Cube for Debut"

Hybrid Memory Cube (HMC) technology, a multi-chip module (MCM) that aims to address one of the biggest challenges in high performance computing: scaling the memory wall.

Memory architectures haven't kept pace with the bandwidth requirements of multicore processors. As microprocessor speeds out-accelerated DRAM memory speeds, a bottleneck developed that is referred to as the memory wall. Stacked memory applications, however, enable higher memory bandwidth.

The Hybrid Memory Cube (HMC) is a new memory architecture that combines a high-speed logic layer with a stack of through-silicon-via (TSV) bonded memory die that enables impressive advantages over current technology. According to company figures, a single HMC offers a 15x performance increase and uses 70 percent less energy per bit when compared to DDR3 memory, and takes up 90 percent less space than today's RDIMMs. The Cube is also scalable per application, which is not possible with DDR3 and DDR4. System designers have the option of employing the HMC as near memory for best performance or in a scalable module form factor, as far memory, for optimum power efficiency. This is a huge leap forward from a technology perspective, noted Graham, compared to DDRx and other boutique memory products that are out there.

http://www.hpcwire.com/hpcwire/2013-...for_debut.html

TheXFiles88

Thanks for the links. Never heard of it till now, so it was interesting to read.

Also, to pour some gasoline on the fire I found this link as well:

It will have applications in servers but the most obvious application is in games. Faster memory together with faster processors will make for smoother high-resolution graphics. That is where you will notice the most change. It won't make any noticeable difference for simple things like accounting or word processing.

Linky

Yeah and this quote:"The Cube is also scalable per application....System designers have the option of employing the HMC as near memory for best performance or in a scalable module form factor, as far memory, for optimum power efficiency"is very interesting since it goes in lieu with that leaked "Scalable Console patent". Here is the chart of how the HMC vs. DRR3 & DDR4:

1

Another stack memory i.e. HBM http://sites.amd.com/us/Documents/TFE2011_006HYN.pdf Document dated 3/10/2011.

http://semiaccurate.com/2011/10/27/amd-far-future-prototype-gpu-pictured/

AMD_Interposer_SemiAccurate.jpg

http://www.eetimes.com/electronics-news/4369672/Micron-broadening-support-for-memory-cube

Micron is not alone in the pursuit of a 3-D memory stack. The Jedec group is working on a follow on to the 12.8 Gbit/second Wide I/O interface that targets mobile applications processors. The so-called HB-DRAM or HBM effort is said to target a 120-128 Gbyte/second interface and is led by the Jedec JC-42 committee including representatives from Hynix and other companies.

AMD would need stack memory for AMD Kaveri APU i.e. running into the memory bandwidth wall with Trinity.

#84 Posted by ronvalencia (15265 posts) -

Sounds like specs got better for Orbis :D

Orbis not only has a more powerful GPU,but now also has 4 compute units dedicated for whatever like cloth physics and such :P

killzowned24

14 + 4 split is most likely using GCN's hardware partitioning feature i.e. a feature not used in DX..

#85 Posted by TheXFiles88 (1040 posts) -

[QUOTE="TheXFiles88"]

[QUOTE="SUD123456"]

Thanks for the links. Never heard of it till now, so it was interesting to read.

Also, to pour some gasoline on the fire I found this link as well:

It will have applications in servers but the most obvious application is in games. Faster memory together with faster processors will make for smoother high-resolution graphics. That is where you will notice the most change. It won't make any noticeable difference for simple things like accounting or word processing.

Linky

ronvalencia

Yeah and this quote:"The Cube is also scalable per application....System designers have the option of employing the HMC as near memory for best performance or in a scalable module form factor, as far memory, for optimum power efficiency"is very interesting since it goes in lieu with that leaked "Scalable Console patent". Here is the chart of how the HMC vs. DRR3 & DDR4:

1

Another stack memory i.e. HBM http://sites.amd.com/us/Documents/TFE2011_006HYN.pdf Document dated 3/10/2011.

http://semiaccurate.com/2011/10/27/amd-far-future-prototype-gpu-pictured/

AMD_Interposer_SemiAccurate.jpg

http://www.eetimes.com/electronics-news/4369672/Micron-broadening-support-for-memory-cube

Micron is not alone in the pursuit of a 3-D memory stack. The Jedec group is working on a follow on to the 12.8 Gbit/second Wide I/O interface that targets mobile applications processors. The so-called HB-DRAM or HBM effort is said to target a 120-128 Gbyte/second interface and is led by the Jedec JC-42 committee including representatives from Hynix and other companies.

AMD would need stack memory for AMD Kaveri APU i.e. running into the memory bandwidth wall with Trinity.

I recognized AMD also has a similar design for its new GPU architectures. However, I am more interested of its "scalability per application" such as that leaked "scalable console patent" from the new Xbox's road map.

#86 Posted by ronvalencia (15265 posts) -

These numbers don't add up for me. 1.6 billion triangles per second seems low while xenon did 500 million per sec. Sony claimed the ps3 could do 2.0 teraflops and this is 1.8? Would it really be that expensive to overclock that CPU to atleast 2.0ghz? Audio processor? Doesn't the CPU usually handle that. Where is ronvalencia when u need him.GotNugz

AMD Xenos couldn't sustain 500 million triangles per second ie.g. it needs GCN's geometry design fixes.

http://www.digitalartsonline.co.uk/news/creative-hardware/siggraph-2012-amd-launches-firepro-w5000-w7000-w8000-w9000-graphics-card/

AMD FirePro W5000 has 1.65 billion triangles per second and it's based on Radeon HD 7850 with 768 stream processors (12 CU).

PS4's Liverpood GPU seems to be GCN v2.0's CU design with 7850's front-end and back-end(1) i.e. it's semi-customised.

1. http://www.amd.com/us/products/desktop/graphics/7000/7850/Pages/radeon-7850.aspx#3

  • 64 Texture Units
  • 128 Z/Stencil ROP Units
  • 32 Color ROP Units
#87 Posted by campzor (34932 posts) -
im still waiting for the dbz power level translation.
#88 Posted by ronvalencia (15265 posts) -

[QUOTE="ronvalencia"]

[QUOTE="TheXFiles88"]

Yeah and this quote:"The Cube is also scalable per application....System designers have the option of employing the HMC as near memory for best performance or in a scalable module form factor, as far memory, for optimum power efficiency"is very interesting since it goes in lieu with that leaked "Scalable Console patent". Here is the chart of how the HMC vs. DRR3 & DDR4:

1

TheXFiles88

Another stack memory i.e. HBM http://sites.amd.com/us/Documents/TFE2011_006HYN.pdf Document dated 3/10/2011.

http://semiaccurate.com/2011/10/27/amd-far-future-prototype-gpu-pictured/

AMD_Interposer_SemiAccurate.jpg

http://www.eetimes.com/electronics-news/4369672/Micron-broadening-support-for-memory-cube

Micron is not alone in the pursuit of a 3-D memory stack. The Jedec group is working on a follow on to the 12.8 Gbit/second Wide I/O interface that targets mobile applications processors. The so-called HB-DRAM or HBM effort is said to target a 120-128 Gbyte/second interface and is led by the Jedec JC-42 committee including representatives from Hynix and other companies.

AMD would need stack memory for AMD Kaveri APU i.e. running into the memory bandwidth wall with Trinity.

I recognized AMD also has a similar design for its new GPU architectures. However, I am more interested of its "scalability per application" such as that leaked "scalable console patent" from the new Xbox's road map.

If Microsoft funds HMC for AMD GPUs, then AMD could benefit from it e.g. AMD's unified shader tech was partly funded by MS.
#89 Posted by clyde46 (46997 posts) -
im still waiting for the dbz power level translation.campzor
Its not over 9000. Yet.
#90 Posted by GotNugz (681 posts) -
[QUOTE="campzor"]I cant understand tech mumbo jumbo.. just tell me with dbz power levels how it compares to durango (rumored) + wii u + ps3/360

Playstation 4 = Goku Xbox 720 = Vegeta Wii U = Krillin PS3/360 = Saibamen
#91 Posted by campzor (34932 posts) -
[QUOTE="GotNugz"][QUOTE="campzor"]I cant understand tech mumbo jumbo.. just tell me with dbz power levels how it compares to durango (rumored) + wii u + ps3/360

Playstation 4 = Goku Xbox 720 = Vegeta Wii U = Krillin PS3/360 = Saibamen

im literally lol'n with the ps3/360 - wii u comparison..
#92 Posted by ronvalencia (15265 posts) -

Uncharted 3 @ 60 fps http://www.eurogamer.net/articles/digitalfoundry-what-if-uncharted-3-ran-at-60-fps

There are some indications Uncharted 3 @60 fps version could be running on a dev PC.

BC (via small exe replacement download) shouldn't be problem with Sony's 1st party games on AMD "Gaming Evolved" based PC box.

#93 Posted by Kingpin0114 (2602 posts) -

[QUOTE="GotNugz"][QUOTE="campzor"]I cant understand tech mumbo jumbo.. just tell me with dbz power levels how it compares to durango (rumored) + wii u + ps3/360campzor
Playstation 4 = Goku Xbox 720 = Vegeta Wii U = Krillin PS3/360 = Saibamen

im literally lol'n with the ps3/360 - wii u comparison..

Yea that was pretty good lol

#94 Posted by ferret-gamer (17352 posts) -

Uncharted 3 @ 60 fps http://www.eurogamer.net/articles/digitalfoundry-what-if-uncharted-3-ran-at-60-fps

There are some indications Uncharted 3 @60 fps version could be running on a dev PC.

BC (via small exe replacement download) shouldn't be problem with Sony's 1st party games on AMD "Gaming Evolved" based PC box.

ronvalencia
Or they just rendered it out at 60fps from a replay like your link implied.
#95 Posted by clyde46 (46997 posts) -
[QUOTE="ronvalencia"]

Uncharted 3 @ 60 fps http://www.eurogamer.net/articles/digitalfoundry-what-if-uncharted-3-ran-at-60-fps

There are some indications Uncharted 3 @60 fps version could be running on a dev PC.

BC (via small exe replacement download) shouldn't be problem with Sony's 1st party games on AMD "Gaming Evolved" based PC box.

ferret-gamer
Or they just rendered it out at 60fps from a replay like your link implied.

If its captured at 30FPS, rendering it again at 60 is going to do nothing.
#96 Posted by rilpas (8222 posts) -
so given the standard rumoured PS4 and Xbox 3 specs, which console seems more powerful?
#97 Posted by tormentos (19277 posts) -
[QUOTE="Tessellation"] we don't know anything about either console,these are just rumors for all we know these two speeds can work together,maybe not? or MS worked it around to fix the bandwidth,remember how they increased the xbox 360 RAM quite late,we don't know anything to base things of just speculations.

The Ram expansion did not fix the xbox 360 bandwidth problems it was the Edram which did it,but it also has a cost to it,because there is a limit on what you can fit on it,the ram addition just helped games by delivering more ram,but without the Edram bandwidth on 360 would have been very starved.
#98 Posted by tormentos (19277 posts) -

[QUOTE="tormentos"][QUOTE="Tessellation"] he was celebrating the 192GB/s of bandwidth and now that they changed it to 176GB/s is around durango's territory,of course he is going to deny these rumors because they don't make his future plastic girlfriend any better :cool:TheXFiles88

Actually i wasn't celebrating but if you most know the PS4 one still quite faster,the 720 on is 68GB/s + 102GB/s but that doesn't mean they should be accounted like 170GB/s,go search on Beyond or Neogaf you can't combine those numbers,is 68 from one point to another and 102 from one point to another.. Unlike the PS4 176GB/s one.

You know, the DDRx DRAM has already been outclassed by a new revolutionary DRAM which called "Hybrid Memory Cube" It is a multi-chip module which works in parallel with the multi-core processors. The advantages are enormous over the DDRx. It costs less, 15x more performance, consumes less power and takes less space than a typical DDR3. If Microsoft is going to use this new HMC in its new Xbox (I don't think why they shouldn't), it will be a game changer since Microsoft is currently 1 of the 10 members of the HMC consortium. Just wait for the final specs after E3...;)

"Hybrid Memory Cube is a revolutionary innovation in DRAM memory architecture that sets a new standard for memory performance, power consumption and cost."

- HMC Combines high-speed logic process technology with a stack of through-silicon-via (TSV) bonded memory die.

- HMC delivers dramatic improvements in performance, breaking through the memory wall and enabling dramatic performance and bandwidth improvements -a single HMC can provide more than 15x the performance of a DDR3 module.

- The revolutionary architecture of HMC is exponentially more efficient than current memory, utilizing 70% less energy per bit than DDR3 DRAM technologies..

- Hybrid Memory Cube's increased density per bit and reduced form factor contribute to lower total cost of ownership, by allowing more memory into each machine and using nearly 90% less space than today's RDIMMs.

- With performance levels that break through the memory wall, Hybrid Memory Cube represents the key to extending network system performance to push through the challenges of new 100G and 400G infrastructure growth. Eventually, HMC will drive exascale CPU system performance growth for next generation HPC systems.

- Whereas DDR4 represents an evolutionary standard, HMC is a revolutionary technology that is a complete paradigm shift from current memory architectures.

- Hybrid Memory Cube will redefine memory. By advancing past the traditional DRAM system, HMC is setting a new standard of memory that can keep up with the advancements of CPUs and GPUs.

- Hybrid Memory Cube could be an absolute game changer for applications ranging from high performance computing to consumer technologies like tablets and graphics cards that value a combination of form factor, energy and bandwidth.

http://www.hybridmemorycube.org/technology.html

"Micron Readies Hybrid Memory Cube for Debut"

Hybrid Memory Cube (HMC) technology, a multi-chip module (MCM) that aims to address one of the biggest challenges in high performance computing: scaling the memory wall.

Memory architectures haven't kept pace with the bandwidth requirements of multicore processors. As microprocessor speeds out-accelerated DRAM memory speeds, a bottleneck developed that is referred to as the memory wall. Stacked memory applications, however, enable higher memory bandwidth.

The Hybrid Memory Cube (HMC) is a new memory architecture that combines a high-speed logic layer with a stack of through-silicon-via (TSV) bonded memory die that enables impressive advantages over current technology. According to company figures, a single HMC offers a 15x performance increase and uses 70 percent less energy per bit when compared to DDR3 memory, and takes up 90 percent less space than today's RDIMMs. The Cube is also scalable per application, which is not possible with DDR3 and DDR4. System designers have the option of employing the HMC as near memory for best performance or in a scalable module form factor, as far memory, for optimum power efficiency. This is a huge leap forward from a technology perspective, noted Graham, compared to DDRx and other boutique memory products that are out there.

http://www.hpcwire.com/hpcwire/2013-...for_debut.html

Shoud i do that a go with GDDR5 best and fastest solution.? No that would be call wishful thinking,the fact that the ESRAM exist basically confirms that you will not get that Ram on 720,ESRAM is there for a reason,to help the starved DDR3 memory.
#99 Posted by ferret-gamer (17352 posts) -

[QUOTE="ferret-gamer"][QUOTE="ronvalencia"]

Uncharted 3 @ 60 fps http://www.eurogamer.net/articles/digitalfoundry-what-if-uncharted-3-ran-at-60-fps

There are some indications Uncharted 3 @60 fps version could be running on a dev PC.

BC (via small exe replacement download) shouldn't be problem with Sony's 1st party games on AMD "Gaming Evolved" based PC box.

clyde46

Or they just rendered it out at 60fps from a replay like your link implied.

If its captured at 30FPS, rendering it again at 60 is going to do nothing.

Uncharted 3 has a replay editor, it wouldn't be that hard for them render it out at 60fps, any decent replay system in a game could do it. I'm pretty sure there is a youtube video of someone who used Source engine's replay system to make a slow mo 10,000fps vid.

#100 Posted by CwlHeddwyn (5314 posts) -
It makes no sense for MS to use DDR3 RAM, it's too slow and will be a handicap.