DDR5-2018?

[iggybo]

Hey guys,

 

What do you think about DDR5 coming to Mobos in 2018? Will you be waiting for your new build until DDR5 is a thing?

Allegedly it boasts with double the speed of DDR4, so for some folks out there it might even make a significant difference...

I'm just curious what you guys are thinking about this.

 

Thx

[Guest]

New mobo, new cpu, new ram, too much to wait imo for an extra 1fps increase

[TheNaitsyrk]

Just now, iggybo said:

Hey guys,

 

What do you think about DDR5 coming to Mobos in 2018? Will you be waiting for your new build until DDR5 is a thing?

Allegedly it boasts with double the speed of DDR4, so for some folks out there it might even make a significant difference...

I'm just curious what you guys are thinking about this.

 

Thx

Probably not? DDR4 is literally fresh, so I don't get why would they release it so soon.

[iggybo]

1 minute ago, TheNaitsyrk said:

Probably not? DDR4 is literally fresh, so I don't get why would they release it so soon.

https://arstechnica.com/gadgets/2017/03/next-generation-ddr5-ram-will-double-the-speed-of-ddr4-in-2018/

[RadiatingLight]

Yeah, and carbon nanotubes will finally be available to consumers in 2015 2016 2017 2018, right?

 

 

[ARikozuM]

I think this is more "we've got tests going" and "we'll announce results in 2018". DDR3 was the main RAM for about 6 years.

[Droidbot]

Just now, iggybo said:

https://arstechnica.com/gadgets/2017/03/next-generation-ddr5-ram-will-double-the-speed-of-ddr4-in-2018/

FINALIZED IN 2018.. DDR4 was finalized in 2012 but only released into the mainstream in 2015. 

[2FA]

DDR5 isn't going to be a physical thing on the market until at least 2020 or 2021.

[doomsriker]

since the switch from DDR3 to DDR4 did nothing for game performance, my crystal ball predicts that trend will continue.

 

 

[iggybo]

2 minutes ago, Droidbot said:

FINALIZED IN 2018.. DDR4 was finalized in 2012 but only released into the mainstream in 2015. 

Who even reads articles?

Sorry, should've read the whole thing...

But still, twice the speed...that's quite a thing

Though if you consider Moores Law

[RadiatingLight]

1 minute ago, doomsriker said:

since the switch from DDR3 to DDR4 did nothing for game performance, my crystal ball predicts that trend will continue.

Well, that was because the speeds were the same in the comparison.

DDR4 has a faster speed than DDR3, which matters for gaming, but when both DDR3 and DDR4 are running at the same speeds, there's almost no difference.

[WMGroomAK]

7 minutes ago, iggybo said:

Hey guys,

 

What do you think about DDR5 coming to Mobos in 2018? Will you be waiting for your new build until DDR5 is a thing?

Allegedly it boasts with double the speed of DDR4, so for some folks out there it might even make a significant difference...

I'm just curious what you guys are thinking about this.

 

Thx

I think Wendell from Level1Techs on their weekly news presented it best and mentioned an upcoming article they are doing on how it might be a good thing but they need to work on more than just speed and decrease the latency.  I think he also mentioned that while the standard will be finalized in 2018, you won't see system support for at least a year or two beyond that as they would still need to have production of the physical chips and have Mobos, chipsets and CPUs that support it.  I think this is part of the reason that AMD is planning for their chipset & Mobos to be standardized until 2020 is to allow for implementation of DDR5 and the next PCIe standard.

[tp95112]

It took forever to adopt ddr4. Am4 for fucksakes just added it to there platform which is exspected to be here till 2020. 

[RadiatingLight]

Just now, iggybo said:

But still, twice the speed...that's quite a thing

Well, you know that intel's CPUs increase performance by 15% every year, right??\

My point is, never take statistics from the manufacturer of a product as fact.

[DocSwag]

My guess is the first ddr5 platforms will appear in 2020 or 2021... Which is perfect because that's when I go to college and so I can build a new pc with ddr5  

[DocSwag]

Just now, RadiatingLight said:

Well, you know that intel's CPUs increase performance by 15% every year, right??\

My point is, never take statistics from the manufacturer of a product as fact.

They're referring to base speed, as I believe ddr3 launched at 1066 MHz and ddr4 at 2133. So ddr4 is double. Chances are ddr5 will start at 4266 MHz.

 

But the thing is, was ddr4 considered double ddr3 at launch? No, because by then ddr3 was selling at 1600-2400 MHz, and ddr4 only at 2133-3000 MHz, so in the end it was more like a 25-50% increase.

 

My guess is by the time ddr5 is available in systems, most ddr4 will be between 3000 and 4000 something MHz, so even then it won't be double.

[Shiv78]

16 minutes ago, iggybo said:

Sorry, should've read the whole thing...

 

Yes, you should have, you posted it. Who posts things without even reading them?

 

Also: 

Quote

Like DDR4 back when it was announced, it will still be several years before any of us have DDR5 RAM in our systems. That's partly because the memory controllers in processors and SoCs need to be updated to support DDR5, and these chips normally take two or three years to design from start to finish. DDR4 RAM was finalized in 2012, but it didn't begin to go mainstream until 2015 when consumer processors from Intel and others added support for it.

 

[JoostinOnline]

32 minutes ago, doomsriker said:

since the switch from DDR3 to DDR4 did nothing for game performance, my crystal ball predicts that trend will continue.

 

 

It did nothing when they were at the same frequency.  In CPU bound situations, higher clocked RAM helps.  DDR4 can be clocked at well over 4000MHz.

 

My DDR3 is clocked at 2400MHz, so it's basically the same as low end DDR4.

[doomsriker]

2 minutes ago, JoostinOnline said:

It did nothing when they were at the same frequency.  In CPU bound situations, higher clocked RAM helps.

 

My DDR3 is clocked at 2400MHz, so it's basically the same as low end DDR4.

actually, Memory bandwidth doesn't matter much either in games

 

[JoostinOnline]

15 minutes ago, doomsriker said:

actually, Memory bandwidth doesn't matter much either in games

 

A common misconception.  It makes a bigger difference the higher you go up, and even then only in CPU bound situations.  You shouldn't believe something just because Linus said it.  He's human and makes mistakes like the rest of us.

As you can see, it makes a huge difference in some situations.

[sgloux3470]

The standard is always finalized long before it hits the market.

 

JEDEC DDR4 was published in 2012 for example, but wasn't really seen until X99 in late 2014.

[ARikozuM]

36 minutes ago, JoostinOnline said:

A common misconception.  It makes a bigger difference the higher you go up, and even then only in CPU bound situations.

 

It makes a difference even when the GPU becomes the bottleneck by allowing better frame times.

[sgloux3470]

35 minutes ago, JoostinOnline said:

A common misconception.  It makes a bigger difference the higher you go up, and even then only in CPU bound situations.  You shouldn't believe something just because Linus said it.  He's human and makes mistakes like the rest of us.

As you can see, it makes a huge difference in some situations.

I think the reason for this perception comes from people not understanding CPU bottlenecks.

 

For example people always claimed that Crossfire and SLI had poorly scaling (50-60%) but if you actually put them under an intense GPU bound scenario you absolutely do see 90-100% scaling. (Like 4K)  The problem is traditionally GPU's that powerful were bottlenecked by the CPU.  It's only recently that single GPU's have gotten powerful enough to start bottlenecking CPU's at 1080p which is why we seem faster RAM becoming so beneficial.  These days multiGPU, especially from AMD, is almost exactly double the performance.

 

In the last couple of weeks I've seen many common misconceptions brought to light.  Ram speeds, Nvidia and AMD driver overhead, i5 versus i7 in gaming.... there's a ton of misinfo out there from shitty test methodology and poorly researched benchmarks.  For example you can see in Linus's video they put a ton of AA on the GPU to stress it, creating a GPU bottleneck when faster memory benefits the CPU. 

 

 

 

[BuckGup]

More like 2020.

[iggybo]

23 hours ago, Shiv78 said:

Who posts things without even reading them?

Meeeee!

And I'm sorry, but I did actually read some other articles but then found arstechnica to be a more credible looking source without having read it. I thought it would basically say the same.

 

Shame on me