One of the problems with high performance NIC's is that the modern PC architecture has a bit of trouble keeping up. But, in your case, this isn't so much the problem. Let me explain.
The CPU has to do a lot of work processing TCP packets. This affects the throughput. What's limiting things in your case is not the network hardware, but the ability of the server to saturate the network links.
In more recent times, we've seen processing move from the CPU to the NIC like checksum offload. Intel have also added features to help reduce the load further. That's cool and I'm sure all optimizing features are turned on.
As you've alluded to, jumbo frames - actually that helps throughput somewhat. But not as much as RDMA.
Most 10GBit ethernet hardware will have a very nice underutilized feature called RDMA or remote direct memory access. It allows the NIC to do memory to memory copies over the network without the intervention of the CPU. Well, OK the CPU tells the NIC what to do and then the NIC does the rest. The trouble is, it's not used much yet. But it's getting there. Apparently, in the most recent version of Microsoft Windows Server 2012, they have something called SMB Direct. It uses RDMA. So, if you want to increase throughput, you want to use that.
Are you able to put together some test hardware and install it onto there to see how it performs?
By the way, I'm not sure if you will see it at 10Gbit so much, but fast RAM helps with RDMA especially with 56Gbit Infiniband. In general, it's best to use the fastest RAM your server supports.
Also note this comment on the SMB Direct link I put above:
You should not team RDMA-capable network adapters if you intend to use
the RDMA capability of the network adapters. When teamed, the network
adapters will not support RDMA.
Update: Looks like not ALL 10GBit NIC's support RDMA for some reason.
So check your model's features first.
Another thought I had was the type of protocol being used to do your testing may be affecting the results. i.e. protocol overhead on top of TCP overhead. I suggest you look into using something that can test without touching the hard drive such as iperf. There is a windows port of it somewhere.
Best Answer
According to this link: http://www.intel.com/support/network/adapter/pro100/sb/CS-032498.htm
These are server based nic's. Having said that, some desktop systems use the the igb driver as well. The 82576 is hardly aimed at desktop use. It's stated to be for dual port NICs and virtualization. What NIC model/chipset do you have?
Edit: You've stated you have the 82576.
I'd expect better. I think the problem is likely the igb driver. Especially if you don't see issues with these cards/nic chipsets under windows.
I suspect he reason the driver is so terribly old is that it's tied to an the kernel version that ESXi is based on. It would be a similar story to XenServer. Even the latest version of XenServer is using a 2.6 series and a distro based on Centos 5.6. It's horrendously old but well tested. There's a good bet this is the reason.
XenServer provide a DDK to allow you to insert other kernel drivers. Do VMWare provide the same sort of facility?Perhaps the update provided by VMWare may not be available in the normal update channel because it's not been thoroughly tested? After all, what they are providing is a backport of the driver to the specific kernel they are running with ESXi.
Reading through some of the release notes of the igb driver and there were numberous bug fixes. Some of them affecting performance. Some fixing broken features. Some enabling features that are turned off by default.
By the way, in answering "How common are intel igb NIC vs e1000 etc? Very common. I've even seen them on some reasonable desktop motherboards.