Hi all,

Today I visited Routerboard.com looking for some performance data about CCR1009-8G-1S-1S+ that I needed to remember, and what was my surprise when I saw that performance numbers had changed since last time I checked them.
I had saved this data some time ago:
( 1518 byte )
Bridging Fast Path: 11,841 Mbps
Routing 25 Rules:   9,882 Mbps

But the performance numbers that the site shows today ( July 7th 2016 ) are:
( 1518 byte )
Bridging Fast Path: 5,914.1 Mbps
Routing 25 Rules:   5,914.1 Mbps

These are far less from what I had, and make no sense since the block diagram shows 5 ethernet lines ( 1 shared among 4 ethernet ports ) conected to the CPU plus 1 SFP and 1 SFP+, that sum up about 15,000 Mbps maximum throughput, that is consistent with the numbers I had ( about 12,000 Mbps for bridging fast path ).
Also, I found almost the same performance numbers in two websites: http://www.meconet.de/cms/upload/public ... ews_57.pdf and 
http://www.xconcepts-me.com/en/productD ... e=1&Next=1.

So, what's the truth? Routerboard.com data is wrong or there has been some hardware changes to the CCR1009 that has limited its throughput?

Thanks in advance

Manuel 

CCR1009 has 4 models:

CCR1009-8G-1S-PC           passive cooling 1.0ghz CPU no SFP+ port 1gb ram, theoretical max possible throughput almost 6gbps
CCR1009-8G-1S-1S+PC     passive cooling 1.0ghz CPU   1 SFP+ port 2gb ram, theoretical max possible throughput almost 16gbps
CCR1009-8G-1S                 active cooling    1.2ghz CPU no SFP+ port 1gb ram, theoretical max possible throughput almost 6gbps (ARCHIVED MODEL)
CCR1009-8G-1S-1S+          active cooling    1.2ghz CPU  1 SFP+  port 2gb ram, theoretical max possible throughput almost 16gbps

performance data on routerboard.com about this 4 models are mixed between brochures and web page and difficult to establish, some times they refer the 1.2ghz cpu performance and sometimes 1.0ghz performance, sometimes SFP+ results and sometimes no SFP+ results

I think almost 6gps on fast path 1518 byte test is consistent with no SFP+ devices
i thing almost 12gbps on fast path 1518 byte test is consistent with SFP+ devices
I think 739.7kpps on 25 ip filter rules 64 byte test is consistent with 1.0ghz cpu clock
I think 878.7kpps on 25 ip filter rules 64 byte test is consistent with 1.2ghz cpu clock

Thanks Chechito for your answer.
I am referring specifically to the CCR1009-8G-1S-1S+ model. This is what is shown in its page at Routerboard.com:

Performance test results
CCR1009-8G-1S-1S+ Tile 9 core all port test
Mode Configuration: 1518 byte

Bridging none (fast path)        487 kpps     5914.1 Mbps
Bridging 25 bridge filter rules  487 kbps     5914.1 Mbps
Routing none (fast path)         487 kpps     5914.1 Mbps
Routing 25 simple queues      487 kpps     5914.1 Mbps
Routing 25 ip filter rules          487 kpps     5914.1 Mbps

And this is what I had and what I have found in other websites with the previously documented performance:

Bridging none (fast path)        975.1 kpps    11,841.6 Mbps
Bridging 25 bridge filter rules  975.1 kpps    11,841.6 Mbps
Routing none (fast path)         975.1 kpps    11,841.6 Mbps
Routing 25 simple queues      975.1 kpps    11,841.6 Mbps
Routing 25 ip filter rules          648.2 kpps      7,871.7 Mbps

Today, I checked the performance data for the CCR1009-8G-1S ( no 10 Gbps interface ) in its page in Routerboard.com, and have found that it's the same as CCR1009-8G-1S-1S+ !...so maybe just a publishing error? 
Please, any Mikrotik engineer could clarify this issue?

Thanks

SFP+ port does not make router faster, CPU is the same on both routers with and without SFP+

SFP+ port does not make router faster, CPU is the same on both routers with and without SFP+

That's correct, but the numbers for SFP+ models still look incorrect: 487 kbps is way too low for CCR1009 with FastPath while 5914.1 Mbps is where six 1G ports are maxed out, and additional 10G port should make that limit bigger. So 11841.6 Mbps with FastPath for 1500 byte packets on SFP+ models looks more credible.

The SFP+ model has one additional (10G) port connected to the CPU directly. It is a question of number of ports.
I see that we accidentally posted wrong results for SFP+ model. We will fix it today. 

Thanks Normis for the clarification!
I have seen today in CCR1009-8G-1S-1S+´s web page that performance numbers are correct now.