News Intel 13th-Gen Raptor Lake Specs, Rumored Release Date, Benchmarks, and More

[Admin]

Here's the Intel 13th-Gen Raptor Lake release , benchmarks, specifications, pricing, and all we know about the new chip architecture.

Intel 13th-Gen Raptor Lake Specs, Rumored Release Date, Benchmarks, and More : Read more

 

[-Fran-]

I think there's a typo/error in the PCIe lanes for the chipsets table where you state: "In the past, the PCH supported up to 16 PCIe 3.0 lanes and up to 12 PCIe 4.0 lanes, but Intel has increased the number of PCIe 4.0 lanes to 20 and reduced the number of PCIe 3.0 lanes to eight", as the table reflect the reverse and had me scratching my head for a bit xD

Regards.

 

[thisisaname]

Even though Raptor Lake is clearly on the cusp of coming to market this year — we've even seen chips sold at auction and benchmarks in the wild — Intel has been uncharacteristically silent about its pending line of chips for desktop PCs.

Sounds to me they are playing a high tech version of Chicken with AMD to see who will blink first and release some real data about their chips.

 

[SiliconFly]

Dear Tom's Hardware,
You have used the word "Rocket Lake" instead of "Raptor Lake" in 11 places in the article. Esp. in the section titles. Kindly fix it as it's a bit misleading.
-SiFy.

 

[jp7189]

I think you should have used a WDC Raptor X . That was a unique and very cool drive. I keep one in a drawer so I can one day show grandkids how a hard drive worked. Lol.

 

[jp7189]

I think you should have used a WDC Raptor X . That was a unique and very cool drive. I keep one in a drawer so I can one day show grandkids how a hard drive worked. Lol.

(In the background of the top pic)

 

[peachpuff]

Fixed ilm?

 

[saltweaver]

Since H710 boards are no more, will Raptor Lake fit on H610 board?

 

[shady28]

A lot of benchmarks show DDR5-6000+ is where DDR5 starts to overtake DDR4 so we may start to see some sizeable performance increases if Intel can get it certified for DDR5-5600, which hopefully will also translate into higher success rates with OC DDR5. DDR5-6200 seems pretty affordable now as well.

I also like the increased push away from the old USB 3 5Gbps standards. I think these new 7XX motherboards will be more expensive because of that, but that's ok by me if they have better connectivity and IO. IO throughput is one of those areas most sites don't test anymore, but there are big differences between different motherboards in this area (m.2 and USB throughputs can vary wildly).

I don't expect raw performance on most tasks to go up on RL though. 5% maybe on single thread, possibly 15% if you factor in higher clocks, while multi-thread is IMO highly overrated and is more of a benchmark competition than being relevant to most users.

 

[PCWarrior]

Well we typically see ES leaks about upcoming cpus first show up around 6-9 months prior to release and these leaks get more frequent and accurate as time goes by. When the launch date is only 2-3 months away we see multiple leaks about performance and specs and a 1 month earlier we even get pricing information. We even see ES out in the wild being sold 1-3 months prior to launch. To me it is certain that Raptor lake will be launching at most 3 months from now – that is early October. The announcement (and sending samples to the tech press for reviewing) should come in mid/late September.

The same applies to AMD and what worries me is that despite AMD cpus being officially stated to launch in fall 2022 and despite being supposedly expected to launch even earlier than Raptor lake, in September, we don’t have any real leaks unlike Raptor lake. This indicates at the very least a very small (or no) volume of cpus circulating outside AMD labs for testing and the last time this happened, with the 3D-Vcache cpus, it meant a delay. Everyone was talking and expecting these cpus to come somewhere in October-December 2021. But we were getting absolutely no leaks. It turned out they weren’t ready and came out much later (April 2022) and it was just a single SKU. My hunch is that AMD will only be announcing, demonstrating and paper launching their cpus in late September and real availability/launch will come significantly later (in early/mid November or so).

These two generations should be trading blows on multiple fronts. Intel's single/light threaded and per core performance of the p-cores should see a 10-15% increase (combined IPC and frequency gains). Purely multithreaded should increase by 30-45% depending on the application. AMD will see a 25-30% increase in single/light threaded (not enough to match Intel which is already 15-30% ahead with Alderlake) and 30% increase in purely multithreaded (depending on the application the 5950X is already trading blows with the 12900K and a 30% increase will keep AMD in close range to Raptor lake but still likely lose overall).

 

[shady28]

Well we typically see ES leaks about upcoming cpus first show up around 6-9 months prior to release and these leaks get more frequent and accurate as time goes by. When the launch date is only 2-3 months away we see multiple leaks about performance and specs and a 1 month earlier we even get pricing information. We even see ES out in the wild being sold 1-3 months prior to launch. To me it is certain that Raptor lake will be launching at most 3 months from now – that is early October. The announcement (and sending samples to the tech press for reviewing) should come in mid/late September.

The same applies to AMD and what worries me is that despite AMD cpus being officially stated to launch in fall 2022 and despite being supposedly expected to launch even earlier than Raptor lake, in September, we don’t have any real leaks unlike Raptor lake. This indicates at the very least a very small (or no) volume of cpus circulating outside AMD labs for testing and the last time this happened, with the 3D-Vcache cpus, it meant a delay. Everyone was talking and expecting these cpus to come somewhere in October-December 2021. But we were getting absolutely no leaks. It turned out they weren’t ready and came out much later (April 2022) and it was just a single SKU. My hunch is that AMD will only be announcing, demonstrating and paper launching their cpus in late September and real availability/launch will come significantly later (in early/mid November or so).

These two generations should be trading blows on multiple fronts. Intel's single/light threaded and per core performance of the p-cores should see a 10-15% increase (combined IPC and frequency gains). Purely multithreaded should increase by 30-45% depending on the application. AMD will see a 25-30% increase in single/light threaded (not enough to match Intel which is already 15-30% ahead with Alderlake) and 30% increase in purely multithreaded (depending on the application the 5950X is already trading blows with the 12900K and a 30% increase will keep AMD in close range to Raptor lake but still likely lose overall).

I expect Zen 4 launch to be much like Zen 3. It wasn't reported on very much, but AMD only produced approximately 1 million Zen 3 in 2020 Q4 with an October "Launch". Moreover, that number is 'charitable' with some unlikely assumptions.

According to this article the number of x86 CPUs shipped in 2021 was 471 Million. Assuming 2020 was similar, AMD was able to supply something like zero point five (0.5) percent of the Q4 market with Zen 3, and only on enthusiast grade desktop. This is the textbook definition of a paper launch. It wasn't until March / April of 2021 that Zen 3 desktop was at least somewhat available albeit at inflated prices, and by then we had Rocket Lake aplenty and Tiger Lake was actually progressing beyond the original low power SKU launch to the higher power H series.

The issue here is that AMD has limited, pre-scheduled production from TSMC. Given that, they will allocate that production first to that required by contract (PS5, XBox) and then to the most profitable SKUs, the rest get leftovers.

The most profitable SKUs are going to be EPYC, and they still have a ton of demand from consoles.

So based on those facts, I would not expect Zen 4 to be a real obtainable product for 99% of the people until 2023. With Intel, you can count on being able to fairly easily obtain their product 4-6 weeks after launch, and if you really want it you can probably get it for a premium in the first couple of weeks.

 

[-Fran-]

I expect Zen 4 launch to be much like Zen 3. It wasn't reported on very much, but AMD only produced approximately 1 million Zen 3 in 2020 Q4 with an October "Launch". Moreover, that number is 'charitable' with some unlikely assumptions.

According to this article the number of x86 CPUs shipped in 2021 was 471 Million. Assuming 2020 was similar, AMD was able to supply something like zero point five (0.5) percent of the Q4 market with Zen 3, and only on enthusiast grade desktop. This is the textbook definition of a paper launch. It wasn't until March / April of 2021 that Zen 3 desktop was at least somewhat available albeit at inflated prices, and by then we had Rocket Lake aplenty and Tiger Lake was actually progressing beyond the original low power SKU launch to the higher power H series.

The issue here is that AMD has limited, pre-scheduled production from TSMC. Given that, they will allocate that production first to that required by contract (PS5, XBox) and then to the most profitable SKUs, the rest get leftovers.

The most profitable SKUs are going to be EPYC, and they still have a ton of demand from consoles.

So based on those facts, I would not expect Zen 4 to be a real obtainable product for 99% of the people until 2023. With Intel, you can count on being able to fairly easily obtain their product 4-6 weeks after launch, and if you really want it you can probably get it for a premium in the first couple of weeks.

Counter argument: Intel 10K and 11K series were widely available because people didn't want them at their launch price points and were willing to wait for Zen3 products. I do recall very well that Intel in previous launches, specially on new nodes, have always had problems supplying the market on the first few months and that has affected prices for a while. That's just par of the course with any new popular release.

Regards.

 

[shady28]

Counter argument: Intel 10K and 11K series were widely available because people didn't want them at their launch price points and were willing to wait for Zen3 products. I do recall very well that Intel in previous launches, specially on new nodes, have always had problems supplying the market on the first few months and that has affected prices for a while. That's just par of the course with any new popular release.

Regards.

I'm sure Intel could pop out a million Raptor Lake CPUs right now, let them gestate for 3-6 months while they debug the BIOS and so on, sort of like AMD did.

10th Gen was difficult to get any K SKUs for about 6 weeks or so after launch, it was very popular and in fact remains popular especially with gamers. Go look at a chart and see where Zen 2 lands in relation to a 10700K or 10900K. A ton of people bought into the mob psychology surrounding Zen 2. It was and is a poor performer for typical day to day tasks and for gaming. Its only saving grace was multi-core performance, something rarely used by either gamers or normal users.

If you don't believe it, go look up some of the 12900K reviews. Many sites didn't even bother putting Zen 2 on the charts. But 10th gen is still there.

Where's Zen 2???

Where's Zen 2???

This is all because these sites hyped Zen 2 as a great gaming GPU for a year, and then this happened :

 

[-Fran-]

I'm sure Intel could pop out a million Raptor Lake CPUs right now, let them gestate for 3-6 months while they debug the BIOS and so on, sort of like AMD did.

10th Gen was difficult to get any K SKUs for about 6 weeks or so after launch, it was very popular and in fact remains popular especially with gamers. Go look at a chart and see where Zen 2 lands in relation to a 10700K or 10900K. A ton of people bought into the mob psychology surrounding Zen 2. It was and is a poor performer for typical day to day tasks and for gaming. Its only saving grace was multi-core performance, something rarely used by either gamers or normal users.

If you don't believe it, go look up some of the 12900K reviews. Many sites didn't even bother putting Zen 2 on the charts. But 10th gen is still there.

Where's Zen 2???

Where's Zen 2???

This is all because these sites hyped Zen 2 as a great gaming GPU for a year, and then this happened :

I didn't mention Zen2, but since you did: its* appeal was, much like Zen3, the easy upgrade path and it's cheap price compared to the Intel platform alternatives. There's a reason the R5 3600 was so praised when it launched at its price point.

Not* everything is top performance, otherwise we wouldn't be praising the 12100 and 12400, would we?

Regards.

 

[shady28]

I didn't mention Zen2, but since you did: its* appeal was, much like Zen3, the easy upgrade path and it's cheap price compared to the Intel platform alternatives. There's a reason the R5 3600 was so praised when it launched at its price point.

Not* everything is top performance, otherwise we wouldn't be praising the 12100 and 12400, would we?

Regards.

You talked about 10th Gen, which was the competitor to Zen 2. Do you think I only get to mention products you approve of and that fit your narrative?

Few people ever upgrade CPUs, it's a small part of the 20% who DIY their desktop, and desktops are only 30% of the overall PC market.

But even if we want to talk about that tiny sub-section, a 10th gen 10400 buyer could upgrade to 10850K/10900K, or to a 11700K/11900K.

All of these are competitive with current Zen 3 and even when using 3070/3080/3090 GPUs. You might lose 5% FPS for a gamer on 10900K vs Zen 3 but Zen 2, Gen 10s competitor, is not competitive here at all. Which of these again was more future proof?

Speaking of 12400, you do realize AMD only recently released their non-X Zen 3 desktop SKUs? What makes you think they care about cost-conscious enthusiasts? Up until AL ripped into their market share, they were the most ridiculously overpriced CPUs for 18 months, and AMD had no intention of changing that.

 

[-Fran-]

You talked about 10th Gen, which was the competitor to Zen 2. Do you think I only get to mention products you approve of and that fit your narrative?

Few people ever upgrade CPUs, it's a small part of the 20% who DIY their desktop, and desktops are only 30% of the overall PC market.

But even if we want to talk about that tiny sub-section, a 10th gen 10400 buyer could upgrade to 10850K/10900K, or to a 11700K/11900K.

All of these are competitive with current Zen 3 and even when using 3070/3080/3090 GPUs. You might lose 5% FPS for a gamer on 10900K vs Zen 3 but Zen 2, Gen 10s competitor, is not competitive here at all. Which of these again was more future proof?

Speaking of 12400, you do realize AMD only recently released their non-X Zen 3 desktop SKUs? What makes you think they care about cost-conscious enthusiasts? Up until AL ripped into their market share, they were the most ridiculously overpriced CPUs for 18 months, and AMD had no intention of changing that.

Yes, I said 10th and 11th gen had issues to sell due to Ryzen competition and that's why they had less issues in terms of availability. I did not intend ot make any comparison to performance; that was all you moving the topic somewhere else. I don't even understand why you're trying to "hindsight 20/20" me here, as anyone who bought either Zen2, 10th gen or even 11th gen has no issues with acceptable performance and the only ones that have a proper upgrade path are people that actually went with AMD; that's not even up for discussion.

I'll stop here as my point was made.

Regards.

 

[antonis mark]

new socket?!
here goes another motherboard pricey and expensive upgrade...
again...

 

[KyaraM]

new socket?!
here goes another motherboard pricey and expensive upgrade...
again...

What? There is no new socket, Intel literally confirmed Raptor Lake is compatible with the current 600-series boards. There is, however, the new 700-series platform now, with better connectivity. That's it. You can slap a Raptor Lake chip on an Alder Lake board no issue if you really want to. BIOS update needed, of course.

Edit:

The Raptor Lake chips will use the same LGA1700 socket as Alder Lake, meaning they have the same socket and pinout, and both Raptor and Alder will be compatible with both the 600- and 700-series motherboards, providing quite a bit of flexibility for both generations.

 

[thisisaname]

Rumours!! I just wish we had some fact, some nice solid facts.

 

[tamalero]

I call bull on those consumption charts.
With 8 P cores and 16 e Cores.. that thing should consume around threadripper levels...

 

[KyaraM]

I call bull on those consumption charts.
With 8 P cores and 16 e Cores.. that thing should consume around threadripper levels...

Yeah, really inconceivable that they worked on power management. Except they literally mentioned that and already handed in the patent for the concept. And the fact that e-cores aren't quite as hungry as p-cores. Or that most people will never, ever use all cores unless in very specific workloads. Or that power limits exist independently of other metrics like clock speeds; and even when it comes to purely clock speeds, at least my 12700k runs at the same 4.7GHz all-core both when drawing 26W or 130W, and the latter is the maximum due to undervolting, with potential upwards if I decide to overclock. The difference between light and high consumption is the load. And power limits cap consumption, it doesn't matter if you manually apply them or not, that is their function, and if Intel says they want to cap power draw at those numbers, then that's what the official specs are. If YOU want to run the CPU uncapped for the miniscule gains that brings, more power to you, but that doesn't change where Intel sets the limits. It also doesn't change that the gains are so small it's not even worth, but oh well.

 

[SiliconFly]

Rumours!! I just wish we had some fact, some nice solid facts.

Not really. RPL 8P+16E will consume less power than Zen 4 16c

 

[Brian D Smith]

NOT bothering to read this update....until Tom's can somehow note the NEW info. I've lost too much of my life to reading the same things over and over...and over.

 

[ArmourSSD]

I'm finally upgrading from an i5 4670k that I've had since 2016 so this is all very exciting stuff to me. I've waited this long so a couple more months is nothing, and it sounds like the wait is going to be well worth it. The 13900k is what I have my heart set on. C'mon September baby.

 

[jp7189]

Personally I'm not a fan of hybrid on the desktop. I have a 12900k and use it for VR. There are several seemingly independent processes that I run - compositor, steamvr, the actual game, and some helpers like fpsvr. I had cyclical thread latency that would change depending on the order the apps were launched or based on what window was on top. Frametime would vary from 3ms to 40ms in a regular pattern. Pinning affinity for every process was a lot of effort, but simply disabling the E cores fixed all the problems. Now what I have is a very expensive and power thirsty 8 core processor. Its looking like the 13th gen will still only have 8P cores and more E cores... no thank you. I'm looking forward to more details regarding Ryzen 7000 series.